Abstracts of the 23rd Cochrane Colloquium
Transcription
Abstracts of the 23rd Cochrane Colloquium
Abstracts of the 23rd Cochrane Colloquium Supplement 2015 Oral sessions Session 01: Diagnostic test accuracy reviews Session 02: Information retrieval Session 03: Primary studies Session 04: Qualitative evidence Session 05: GRADE Session 06: Disseminating evidence Session 07: Communicating evidence Session 08: Setting priorities to overcome inequality Session 09: Systematic review and meta-analysis methods Session 10: Filtering the information overload for better decisions Session 11: Rapid reviews and overviews of reviews Session 12: Prediction models 1 3 6 44 46 48 76 78 81 117 120 122 Rapid oral sessions Session 01: Risk of bias assessment tools 8 Session 02: Editorial issues 11 Session 03: Advanced methods: Overviews of reviews and network meta-analyses 16 Session 04: Quality of reporting 19 Session 05: Online databases for evidence dissemination 23 Session 06: Individual patient data 26 Session 07: Quality of reporting 50 Session 08: Evidence maps and priority setting 55 Session 09: Meta-analysis methods 59 Session 10: Filtering the information overload for better decisions 62 Session 11: Knowledge translation 67 Session 12: Specific study types / Avoiding research waste 70 Session 13: Translations 83 Session 14: Education and training 88 Session 15: Implementation and qualitative synthesis 91 Session 16: Rapid reviews 95 Session 17: Network meta-analysis 98 Session 18: Outcomes 101 Session 19: Global health: Helping with evidence 125 Session 20: Disseminating evidence 129 Session 21: Systematic review and meta-analysis methods 133 Session 22: Guidelines and GRADE 139 Session 23: Information retrieval 143 Session 24: Novel approaches of interpreting and implementing healthcare evidence 147 Lunch rapid oral sessions Session 01: Minimally important difference Session 02: Priorisation Session 03: Information retrieval Session 04: Decision-making: End users and consumers Session 05: Primary studies Session 06: Health Economics Session 07: Involvement of users and stakeholders Session 08: Search filters Session 09: Trial registration Session 10: Missing data Session 11: Diagnostic test accuracy reviews Session 12: Decision-making: Policy and public health 30 35 37 41 106 109 112 114 152 155 159 161 Poster abstracts Session 1 Session 2 165 224 Workshops Workshops 291 Abstracts available online at www.cochrane.org Abstracts of the 23rd Cochrane Colloquium Oral session 1 Diagnostic test accuracy reviews O 1.1 Authors need guidance on creating clinical pathways in Cochrane diagnostic test accuracy systematic reviews Gopalakrishna G1 , Langendam M1 , Scholten R2 , Bossuyt P1 , Leeflang M1 1 University of Amsterdam, The Netherlands; 2 Dutch Cochrane Centre, The Netherlands Background: Developing the clinical pathway of a medical test can help test accuracy review authors define more precisely the target condition and study eligibility criteria for their reviews. Such pathways also help in understanding the impact of testing on clinical management decisions and ultimately the impact on patient outcomes. Describing the clinical pathway is a recent mandatory requirement in Cochrane diagnostic test accuracy (DTA) reviews but there is no explicit guidance on how to create such a pathway. Objectives: To investigate how authors of Cochrane DTA Review protocols described the clinical pathway in their reviews. Methods: We searched the Cochrane Database of Systematic Reviews for all DTA review protocols published between 2014 and March 2015. We checked if the pathways fulfilled a set of minimum criteria adapted from the analytical pathway as published in the Agency for Healthcare Research and Quality’s medical test manual. Results: We found a total of 31 protocols. All addressed the clinical pathway differently, both in content and format. Most (n = 26) provided only a textual description. Four included both a visual and textual description, and one only a visual description. A detailed analysis of the 10 most recent protocols showed tremendous variation in these pathways (Table 1). The majority did not put the tests in the context of previous or further testing, clinical management following the test(s) or the impact on patient outcomes. In case of single test evaluations, downstream consequences of introducing the test were often not compared to the consequences of not testing. Conclusions: In Cochrane DTA Review protocols there is remarkable variation in interpretation, content and format of the clinical pathway. Review authors could benefit from guidance on how to create pathways to guide them in their evidence selection and ultimately put that evidence in context of downstream patient and/or management consequences. Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. O 1.2 What is the test’s accuracy in my practice population Tailored meta-analysis provides a plausible estimate Willis BH1 , Hyde CJ2 1 University of Birmingham, United Kingdom; 2 University of Exeter, United Kingdom Background: Diagnostic test accuracy studies and metaanalyses may, in some cases, provide estimates that are highly improbable in practice. Tailored meta-analysis provides a potential solution. Objective: To investigate the utility of tailored meta-analysis in synthesizing estimates of a test’s accuracy compared with conventional meta-analysis for three case examples. Method: MEDLINE, EMBASE, and CINAHL for relevant studies; routine data collected on the test positive rate and disease prevalence from the case settings to define an applicable region for each setting. Three cases were evaluated: (1) Mammography in the NHS Breast Screening Program; (2) PHQ-9 questionnaire to screen for depression in general practice; (3) Centor’s criteria used to diagnose group A beta-hemolytic streptococcus in general practice. For conventional meta-analysis, studies were selected using standard systematic review methods; for tailored meta-analysis this selection was refined to those with results compatible with the applicable region for the setting. Baseline study characteristics, quality items and 2 × 2 tables completed. Results: In each example, studies were excluded as a result of incorporating an applicable region for the setting. Comparing tailored with conventional meta-analysis, 1 Cochrane Database Syst Rev Suppl 1–327 (2015) 2 the positive likelihood ratios (with 95% confidence intervals in brackets) were 36.5 (23.0 to 57.9) and 19.8 (12.8 to 30.9) respectively for mammography; and 4.89 (2.02 to 11.8) and 2.35 (1.51 to 3.67) respectively for Centor’s criteria. This had the effect of increasing the positive predictive value from 17% to 27% for mammography and 23% to 38% for Centor’s criteria. Conclusion: Tailored meta-analysis has the potential to provide a plausible estimate for a test’s accuracy which is specific to the practice setting. When compared with conventional meta-analysis, the difference may, in some cases, be sufficient to lead to different decisions on patient management. O 1.3 Including unpublished data in a systematic review of diagnostic test accuracy: impact on summary estimates Korevaar D1 , Westerhof G2 , Wang J1 , Cohen J1 , Spijker R3 , Sterk P2 , Bel E2 , Bossuyt P1 1 Department of Clinical Epidemiology, Biostatistics and Bioinformatics, Academic Medical Center, University of Amsterdam, Netherlands; 2 Department of Respiratory Medicine, Academic Medical Centre, University of Amsterdam, Netherlands; 3 Medical Library, Academic Medical Center, University of Amsterdam, Netherlands Background: Studies with favorable results have higher chances of being published than those with unfavorable results. This could introduce bias in systematic reviews. Objectives: We aimed to compare summary estimates calculated from published and unpublished data in a systematic review of diagnostic test accuracy. Methods: We gathered published data by searching MEDLINE and Embase for studies evaluating the diagnostic accuracy of Fraction of exhaled Nitric Oxide (FeNO), blood eosinophils or Immunoglobulin E (IgE) for detection of airway eosinophilia in adults with asthma (inception to August 2014). We obtained unpublished data by contacting authors of studies that did not report on the diagnostic accuracy of these markers, but seemed to have collected data from which estimates could be calculated, and performed random-effects meta-analysis of area under the receiver operating curve (AUC). Results: Analysing 2919 search results, we included 12 published studies reporting on the accuracy of FeNO, blood eosinophils or IgE (published data). We contacted the authors of two of Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Cochrane Database Syst Rev Suppl 1–327 (2015) 3 O 1.4 providing similar point estimates and confidence intervals of sensitivity and specificity, while prediction intervals were narrower. In Lymphangiography review data, both BM with nominal random effects and BM with latent classes showed a better fit compared to standard BM, providing us with a slightly higher estimate of specificity and, again, with narrower prediction intervals. Conclusions: Although simulation studies are needed, we can conclude that BM with non-parametric random effects could represent a valid alternative to the standard BM to obtain more accurate prediction intervals of sensitivity and specificity. References: Diaz, Mireya. Performance measures of the bivariate random effects model for meta-analyses of diagnostic accuracy. Computational Statistics & Data Analysis 2015: 83: 82–90. Glas, Afina S., et al. Tumor markers in the diagnosis of primary bladder cancer. A systematic review. The Journal of urology 2003: 169.6: 1975–1982. Scheidler, Juergen, et al. Radiological evaluation of lymph node metastases in patients with cervical cancer: a meta-analysis. Jama 1997: 278.13: 1096–1101. A generalized latent variable modelling framework for the bivariate analysis of sensitivity and specificity in diagnostic reviews Oral session 2 Information retrieval these studies, and obtained additional diagnostic accuracy data of markers that had also been assessed, but for which no accuracy results were reported (unpublished data). Authors of another 14 articles were contacted because they could have data from which accuracy estimates could be calculated. Seven shared the requested data, and we also included one conference poster (unpublished data). Adding unpublished data increased sample size for FeNO and blood eosinophils by 28%, and for IgE by 103%. Summary AUC for FeNO was 0.75 (95%CI 0.72 to 0.78; 9 studies; 2263 patients) based on published data, vs 0.73 (0.68 to 0.78; 7 studies; 627 patients) for unpublished data. For blood eosinophils, these numbers were 0.77 (0.71 to 0.83; 7 studies; 1592 patients) vs 0.77 (0.71 to 0.83; 5 studies; 446 patients), and for IgE 0.64 (0.57 to 0.70; 2 studies; 300 patients) vs 0.61 (0.54 to 0.67; 4 studies; 310 patients). Conclusions: Many authors seem to be willing to share unpublished data. Adding unpublished data led to a considerable increase in precision, but barely affected summary estimates of accuracy. Eusebi P1 , Orso M1 , Reitsma JB2 , Vermunt JK3 1 Regional Health Authority of Umbria, Italy; 2 University of Utrecht, The Netherlands; 3 University of Tilburg, The Netherlands Background: The current methodology for diagnostic reviews relies on hierarchical models with normal random effects for the true logit sensitivity and specificity. The assumption of normally distributed random effects can lead to overestimation of between-study variances in the case of outliers, tails and mixtures, and this can results in extremely wide prediction intervals. Furthermore, a considerable number of studies is required to estimate the random effects reliably (Diaz 2015). This results in optimistic confidence intervals and large prediction intervals when substantial heterogeneity is present. Thus it is vital to expand the current methodology to address these issues. Objectives: i) to develop new Bivariate Models (BM) for diagnostic reviews incorporating non-parametric random effects; ii) to provide a comprehensive review of all the methodological extensions of the BMs for diagnostic reviews within the framework of generalized latent variable modelling. Methods: Several specifications of non-parametric random effects will be formulated for the BM. These will include: i) ordinal random effects; ii) nominal random effects; iii) latent classes. Data from two published meta-analyses were analysed for illustrating the properties of the models: ’Telomerase as a diagnostic marker for bladder cancer’ (Glas 2003), and ’Lymphangiography for the diagnosis of lymph node metastasis in women with cervical cancer’ (Scheidler 1997). Results: In Telomerase review data, BM with ordinal random effects showed a better fit compared to standard BM, Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. O 2.1 Improving access to reports of RCTs in Embase: innovative methods enhance the Cochrane Central Register of Controlled Trials (CENTRAL) Glanville J1 , Dooley G2 , Noel-Storr A3 , Foxlee R4 1 York Health Economics Consortium, United Kingdom; 2 Metaxis, United Kingdom; 3 Cochrane Dementia and Cognitive Improvement Group, United Kingdom; 4 Cochrane Editorial Unit, United Kingdom Background: Cochrane Systematic Reviews rely on the efficient identification of research evidence, specifically evidence from randomised controlled trials (RCTs). The largest single source of reports of RCTs is the Cochrane Central Register of Controlled Trials (CENTRAL) in the Cochrane Library. CENTRAL is mainly populated with records from MEDLINE, but also contains a substantial and growing number of records from Embase. The objective was to develop a new bespoke search filter to identify reports of RCTs and novel methods to assess the high volume of candidate reports resulting from the filter. Methods: We developed, validated and refined a sensitive search filter to identify reports of RCTs in Embase. This filter was developed using textual analysis of ten gold standard sets of RCT records (totalling 10,000 records over ten years). The filter performance was tested on a second set of 10,000 RCT reports. Once implemented, records retrieved by the filter were assessed for relevance by a novel crowdsource approach. The search filter was refined after one year of operation based on an assessment of the Cochrane Database Syst Rev Suppl 1–327 (2015) 4 records rejected by the crowd. Results: The development of the search filter and the analysis of output from Embase has resulted in a tiered assessment process, where the most obvious RCT reports are fast-tracked for publication in CENTRAL, leaving more capacity to assess the relevance of less obvious candidate records. Over a 15-month period the filter has identified 198,960 records and 55,042 reports of RCTs have been added to CENTRAL (precision 28%). Conclusions: The records identified by the filter and the crowdsource process have made many thousands of reports of RCTs that were unique to Embase, available in CENTRAL at a high level of precision. These RCTs might be otherwise inaccessible to Cochrane authors since many of them may not have access to Embase. O 2.2 Systematic reviews and meta-analyses of traditional Chinese medicine must search Chinese databases to reduce language bias selection bias. The bias may be particularly severe in systematic reviews of traditional Chinese medicine (TCM) as most randomized controlled trials (RCT) in TCM are published and accessible only in Chinese. Objectives: In this study we investigated how often Chinese databases were not searched in systematic reviews of TCM, how many trials were missed, and whether a bias may occur if Chinese databases were not searched. Methods & Results: We searched five databases in English and three in Chinese for RCTs of Chinese herbal medicine for coronary artery disease and found that 96.64% (115/119) of eligible studies could be identified only from Chinese databases. In a random sample of 80 Cochrane Reviews on TCM, we found that Chinese databases were only searched in 43 (53.75%), in which almost all the included studies were identified from Chinese databases. We also compared systematic reviews of the same topic and found that they may draw a different conclusion if Chinese databases were not searched. Conclusions: In conclusion, an overwhelmingly high percentage of eligible trials on TCM could only be identified in Chinese databases. We suggest that reviewers of TCM search Chinese databases to reduce potential selection bias. Wu X1 , Tang J2 , Mao C2 , Yuan J1 , Qin Y1 , Chung V1 1 The Chinese University of Hong Kong, China; 2 The Hong Kong Branch of the Chinese Cochrane Centre, China Background: Systematic reviews that fail to search non-English databases may miss relevant studies and cause Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Cochrane Database Syst Rev Suppl 1–327 (2015) 5 O 2.3 Web searching for systematic reviews: how to develop the current guidance on methods and reporting standards Briscoe S1 1 University of Exeter, United Kingdom Background: The development of the world wide web has given researchers easy access to information. Although this is a positive development, the web has also contributed to the problem of ’information overload’. This problem may be alleviated through guidance on the most effective methods and reporting standards for web searching. To this end, the author of this abstract has been asked to develop the Cochrane Handbook section on web searching (section 6.2.2.6) for the next major revision. Objectives: This oral session will review the current guidance on web searching for systematic reviews and consider how to develop this guidance. In particular, the session will consider methods for web searching using search engines and websites, and the reporting standards required to ensure transparency and Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. reproducibility of searches. Methods: Current guidance on web searching for systematic reviews will be reviewed, including guidance from Cochrane and the Centre for Reviews and Dissemination (UK). In order to develop this guidance, and searchers’ understanding of web searching, the session will contrast some of the differences between web searching and searching bibliographic databases. This will include the lack of structured indexing on the web by comparison to bibliographic records, and the difference between web-based probabilistic search algorithms and the Boolean searches used by bibliographic databases. Results: It will be demonstrated how the search features of search engines and websites impact on optimal search methods and reporting standards. Recommendations for developing guidelines in line with these findings will be presented. Conclusions: The guidance on web searching to resource systematic reviews currently lacks detail and should be updated. Better guidance on web searching would help searchers and researchers resource systematic reviews and would contribute to ’filtering the information overload’. Cochrane Database Syst Rev Suppl 1–327 (2015) 6 O 2.4 The objectively-derived approach enables the development of high-quality search strategies for bibliographic databases Hausner E1 , Guddat C1 , Hermanns T1 , Lampert U1 , Waffenschmidt S1 1 IQWiG, Germany Background: The development of strategies for systematic searches in bibliographic databases is a major challenge. The search strategies are often complex and may comprise several pages per database searched. The pertinent literature currently recommends a conceptual approach for the development of high-quality search strategies, which is based on a comprehensive collection of synonyms. However, an objectively-derived approach using terms identified in a text analysis may offer advantages. We conducted two projects to evaluate the objectively-derived approach. The first was a retrospective validation of the objectively-derived approach by means of published Cochrane Reviews. The results were promising, so that the conceptual and objectively-derived approaches were prospectively compared in a second project. Objectives: To directly and prospectively compare the objectively-derived approach with the conceptual approach in the development of search strategies. Methods: Search strategies were developed in parallel in MEDLINE for five different research questions: external experts used the conceptual approach to develop the related search strategies; IQWiG used the objectively-derived approach for the same research questions. Results: The research questions of the five projects varied greatly and comprised searches for studies of tyrosine kinase inhibitors, renal cancer therapy, continuous blood-pressure measurement, effectiveness and sustainability of inpatient psychiatric rehabilitation for adults, and impeding and promoting factors in the implementation of clinical guidelines. The objectively-derived approach achieved a weighted mean sensitivity of 97% (95% confidence interval (CI) 94% to 100%) with a weighted mean precision of 5% (95% CI 1% to 10%). The conceptual approach achieved lower values, namely, a weighted mean sensitivity of 75% (95% CI 47% to 100%) and a weighted mean precision of 4% (95% CI 0% to 8%). Conclusions: The objectively-derived approach for the development of search strategies achieves higher sensitivity than the conceptual approach. It produces robust search strategies and in particular offers advantages for complex research questions. Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Oral session 3 Primary studies O 3.1 Olanzapine for schizophrenia: what do the unpublished clinical trials reveal? Lawrence KA1 , Beaumier J2 , Wright J3 , Perry, Jr. T1 , Puil L1 , Turner E4 , Mintzes B5 1 UBC Therapeutics Initiative, Canada; 2 Vancouver Coastal Health Authority, Canada; 3 University of British Columbia, Canada; 4 Oregon Health and Science University, USA; 5 University of Sydney, Australia Background: Selective reporting of data is a major obstacle to advancing and practicing evidence-informed medicine. Even with access to high-quality Cochrane Reviews, the true net benefit versus harm of an intervention cannot be established without inclusion of the data from unpublished clinical trials. Objectives: To compare data on key outcomes from published randomized controlled trials (RCTs) and unpublished clinical trials concerning the antipsychotic olanzapine versus placebo. Methods: Using Cochrane search methods and a trial search co-ordinator, all relevant, published RCTs were retrieved comparing olanzapine with placebo in the treatment of schizophrenia. Through the European Medicines Agency, all relevant clinical study reports submitted by Eli Lilly for olanzapine’s market authorisation for schizophrenia were requested and obtained. We will compare data on the following key outcomes in published and unpublished data, and will examine the effects of any discrepancies on results of meta-analyses: all-cause mortality, non-fatal serious adverse events, and quality of life. Additionally, methods descriptions in published and unpublished trial reports, and assessment of risk of bias, will be compared. Conclusions: Accessing and including data from unpublished clinical trials may prove to create a more accurate assessment of benefit versus harm of a medical intervention. Although the volume of clinical trial data is large and the undertaking complex, this additional level of scrutiny may be required in order to produce the most accurate and unbiased Cochrane Reviews. Cochrane Database Syst Rev Suppl 1–327 (2015) 7 O 3.2 The association between baseline differences and mortality in trials of atypical antipsychotics in dementia: a meta-regression analysis Hulshof T1 1 University Medical Center Groningen, The Netherlands Background: In 2005, a meta-analysis of 15 randomized placebo-controlled trials showed that atypical antipsychotics were associated with an increased risk of mortality in patients with dementia. Consequently, health authorities warned against their use. However, many trials had baseline differences that might explain the meta-analytic findings. Objective: To explore whether the association between baseline differences and mortality could be estimated with a meta-regression analysis. Methods: We abstracted from the articles: (1) randomization features; (2) trial characteristics; and (3) five baseline characteristics that predict death, that is higher mean age, neuropsychiatric symptom (NPS) score, and higher percentage of male, non-white and vascular/mixed dementia patients per treatment group. Each characteristic was scored as high risk if higher or unclear in active treatment group, and otherwise as low. Results: Randomization procedures were poorly described in all trials. Two reports did not present a baseline table. We found not high percentage of males in the active treatment group but studies with a high percentage of females yielded a pooled increased risk of death. The other variables predicted an increased risk of mortality as expected. For each of the five baseline characteristics there were more studies that scored high risk. We then calculated a sum-variable ‘baseline differences’ per study that indicated the number of variables that predicted death (0-5). Both studies that had a haloperidol group in addition to a atypical antipsychotic group scored 0, and nine of the other studies had a score 3 , 4 of 5. In a meta-regression analysis of 12 trials, this sum-variable predicted the risk of mortality (beta 0.32; 95% CI -0.03 to 0.7). Conclusion: Unfavorable baseline characteristics in the atypical antipsychotic group might have mistakenly suggested an increased mortality rate in randomized placebo-controlled trials. O 3.3 Meta-epidemiology of re-analyses of randomized controlled trial data Ebrahim S1 , Sohani Z2 , Montoya L3 , Agarwal A2 , Ebrahim KT2 , Mills E2 , Ioannidis J4 1 McMaster University; Stanford University, Canada; 2 McMaster University, Canada; 3 University Health Network, Canada; 4 Stanford University, USA Background: Reanalyses of randomized controlled trial (RCT) data may help the scientific community assess the validity of reported trial results, but it may also cause confusion. Objectives: To identify reanalyses of RCT data; Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. characterize methodological and other differences between the original trial and re-analysis; evaluate the independence of authors performing the re-analyses; and compare re-analysis findings to those from the original trial to assess whether the re-analysis modified interpretations from the original paper about the kinds or number of patients who should be treated. Methods: We completed an electronic search of MEDLINE from inception to 9 March 2014 to identify all published studies that completed a re-analysis of individual patient data from previously published RCTs addressing the same hypothesis as the original RCT. Four data extractors independently screened papers and extracted data. We assessed for changes in direction and magnitude of treatment effect, statistical significance, and interpretation about who should be treated. Results: We identified 37 eligible reanalyses in 36 published papers, five of which were performed by entirely independent authors, (two based on publicly available data, two on data that were provided on request, the availability of data was not clear for one). Reanalyses differed most commonly in statistical or analytical approaches (n = 18) and in definitions or measurements of the outcome of interest (n = 12). Four reanalyses changed the direction and two changed the magnitude of treatment effect, while four reanalyses led to changes in statistical significance of findings. Thirteen reanalyses (35%) led to interpretations that differed from those of the original paper: three (8%) could be interpreted as showing that different patients should be treated, one (3%) that fewer patients should be treated, and nine (24%) that more patients should be treated. Conclusions: A small number of reanalyses of RCTs have been published to-date and original study conclusions were altered in 35%. O 3.4 Are we overestimating harm when high numbers of placebo patients are switched to active therapy as part of an adaptive trial design? A meta-epidemiological study Tarp S1 , Dossing A1 , Furst DE2 , Boers M3 , Luta G4 , Gluud C5 , Bliddal H1 , Christensen R1 1 Musculoskeletal Statistics Unit, The Parker Institute, Department of Rheumatology, Copenhagen University Hospitals, Bispebjerg and Frederiksberg, Denmark; 2 University of California, Los Angeles, David Geffen School of Medicine, Los Angeles, USA; 3 Department of Epidemiology and Biostatistics, and Amsterdam Rheumatology and Immunology Center at VU University Medical Center, Amsterdam, The Netherlands; 4 Department of Biostatistics, Bioinformatics, and Biomathematics, Georgetown University Medical Center, Washington, DC, USA; 5 Copenhagen Trial Unit, Centre for Clinical Intervention Research, Rigshospitalet, Copenhagen University Hospital, Copenhagen, Denmark Background: Placebo-controlled trials in rheumatoid arthritis (RA) increasingly employ adaptive trial designs that offer early rescue therapy for participants with a poor response at an interim time point. This may compromise inferences drawn from an ITT analysis. Objective: To examine if adaptive trial design confounds inference from ITT analysis Cochrane Database Syst Rev Suppl 1–327 (2015) 8 and to explore the quantified impact of adjusting for skewed dropout between trial groups in the evaluation of harm outcomes. Methods: A systematic search identified RA trials comparing a biologic or tofacitinib vs placebo in approved dose. The main outcome was the ratio of odds ratios (ROR) quantifying the degree of bias associated with adaptive vs non-adaptive trials. An ROR > 1 implies that adaptive trials exaggerate serious adverse event (SAE) estimates. Subsequently, two meta-analyses of trials were performed based on participant number with SAEs (numerator): i) based on the ITT population (denominator) expressed as odds ratio (OR); ii) based on the total person-years of exposure (denominator) expressed as rate ratio (RR). If not reported, exposure was estimated by assuming a linear dropout rate from baseline to follow-up. To compare the pooled OR and RR, the ratio of ratios were calculated as Exp(ln[OR]–ln[RR]), assuming acceptable comparability due to low event rate. A ratio of OR to RR > 1 implies that OR exaggerates harm estimates. Results: All 10 drugs except infliximab and anakinra were represented in both the 69 non-adaptive and the 31 adaptive trials included in the analysis (32,101 patients, approximately 18,902 person-years). Figure 1 illustrates pooled OR for adaptive and non-adaptive trials. The estimate was exaggerated in adaptive trials (ROR 1.29, 1.06 to 1.57). The ratio between the OR and the RR found that OR exaggerates harm estimates (ratio 1.11, 1.08 to 1.14). Conclusion: This study suggests that evaluation of harm outcomes (e.g. SAEs) is affected by adaptive trial designs per se. Applying ITT analysis for odds or risk statistics might overestimate the true effect when analysing harm data from trials with skewed dropout rates. We suggest that an exposure-time adjusted model should be applied. Background: The ACROBAT-NRSi tool is designed to promote rigour in the assessment of risk of bias (RoB) in non-randomised studies (NRS). The development and text of the ACROBAT-NRSi tool draws heavily on clinical examples to illustrate the tool and its application, but the feasibility of it being applied to public health evidence has not yet been fully explored. Objectives: To assess the feasibility and usefulness of applying the ACROBAT-NRSi tool to NRS included in public health intervention reviews. Methods: We applied the ACROBAT-NRSi tool to studies of housing improvement included in a published Cochrane Review. A group of five researchers, all with experience of critical appraisal, read the guidance and met to discuss queries and agree protocol level considerations before assessing the studies independently. Overall assessment of RoB was compared with two tools used to assess the studies in the review (Effective Public Health Practice (EPHP) tool and the Cochrane RoB tool). Results: Agreement in assessments improved over time, but initially considerable discussion was required to agree on the interpretation of criteria and their application to complex social interventions. Specific issues included the incorporation of quasi-experimental study designs and the extent to which confounding may be uncontrollable. The ACROBAT RoB for some studies was ’low’ for two domains (intervention departure and measurement), but the overall assessment was ’serious’ or ’critical’ for all studies. Using the Cochrane RoB, all studies had two or more (range 2 to 8) domains assessed as ’high’, implying a ’high’ RoB overall. The EPHP overall assessment was ’low’ for two studies, and ’moderate’ for three studies. Conclusions: ACROBAT-NRSi may benefit from further development to improve applicability to quasi-experimental studies of social interventions. The tool has limited sensitivity to identify variations in study quality. Where the RoB is ’serious’ or ’critical’ for a body of evidence, greater sensitivity to variations in study quality could allow differentiation of best available evidence from poorer quality evidence. RO 1.2 Assessing the usability of ACROBAT-NRSI for studies of exposure and intervention in environmental health research Rapid oral session 1 Risk of bias assessment tools RO 1.1 ACROBAT-NRSi for public health: reporting on feasibility and utility of applying ACROBAT to studies of housing improvement Thomson H1 , Campbell M2 , Craig P2 , Hilton Boon M2 , Katikireddi V2 1 Cochrane Public Health Group, University of Glasgow, United Kingdom; 2 MRC/CSOSocial&PublicHealthSciencesUnit, University of Glasgow, United Kingdom Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Morgan RL1 , Thayer K2 , Guyatt G1 , Blain R3 , Eftim S3 , Ross P3 , Santesso N1 , Holloway AC4 , Schünemann HJ1 1 Department of Clinical Epidemiology and Biostatistics, McMaster University, Canada; 2 National Toxicology Program Office of Health Assessment and Translation, National Institutes for Environmental Health Sciences, USA; 3 ICF International, USA; 4 Department of Obstetrics and Gynecology, McMaster University, USA Background: Recently, A Cochrane Risk Of Bias Assessment Tool (ACROBAT) was released to examine internal validity (risk of bias) in non-randomized studies of interventions (NRSI). The applicability of ACROBA- NRSI to studies dealing with exposures or interventions in environmental health has not yet been explored. Objectives: This study Cochrane Database Syst Rev Suppl 1–327 (2015) 9 evaluated the application of ACROBAT-NRSI in studies of environmental health exposure. Methods: Two independent reviewers sequentially applied ACROBAT-NRSI to two systematic reviews containing 14 and 17 individual studies of environmental exposures. Material available to reviewers included instructions for application of the ACROBAT-NRSI, including a list of possible confounders specific to the exposures considered in the two reviews. After completing the first and second assessments with the tool, reviewers provided feedback regarding clarity of the instrument and ease of application, highlighting barriers to use. Modifications to enhance usability for studies of exposures (rather than interventions) were made to the tool. Results: Based on feedback from reviewers, two modifications were required: 1. replacement of ’intervention’ with ’exposure’ throughout the document; and 2. provision of additional explanation and examples for application of the tool to studies using a cross-sectional design. Modification to the tool increased reviewers’ understanding during the second application; however, disagreements occurred when judgments on domain-level risk of bias were made, specifically in domains assessing bias due to confounding and bias in the selection of the reported result. Conclusion: Modifications to the tool were limited and improved understanding, as well as reduced disagreement, among reviewers. We identified areas requiring more clarification to further improve consistency of the judgments. We will perform additional work examining studies of exposure with the goal of providing a definitive instrument to assess risk of bias in studies of environmental exposures. RO 1.3 Bias in measurement of adverse outcomes in non-randomised studies Junqueira D1 , Lana LGC1 , Menezes de Pádua CA1 , Zorzela L2 , Vohra S2 1 Universidade Federal de Minas Gerais, Brazil; 2 University of Alberta, Canada Background: The assessment of risk of bias (RoB) in systematic reviews has recently evolved with the development of a Cochrane RoB tool for non-randomised studies of interventions (ACROBAT-NRSI). The tool evaluates internal validity of non-randomised studies (NRS) through signalling questions addressing domains of risk of bias. One of the domains assesses bias in measurement of outcomes, with no distinction between efficacy or harmful outcomes. Objectives: To investigate potential bias with regard to a set of signalling questions proposed to investigate misclassification of adverse outcomes. Methods: We used a sample of NRS from a systematic review of a specific adverse effect (lipodystrophy induced by antiretroviral therapy). The signalling questions included: (1) use of precise definition of the adverse effect, (2) active mode of harms data collection, (3) accurate mode of harms detection, (4) comparability of outcome assessment across intervention groups, (5) specification of Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. who ascertained the adverse effect, (6) training of the person who ascertained the adverse effect, and (7) reporting of timing and frequency of harms data collection. Results: Despite the fact that all studies (n = 20) were designed to investigate the adverse effect lipodystrophy, 25% failed to provide a clear definition for the outcome. Studies did not specify either who ascertained the adverse effect nor the training level of the staff (20% and 25%, respectively), were not designed to actively and accurately detect the adverse effect (15% and 20%), and failed in reporting timing and frequency of harms data collection (15%). Comparability of outcome assessment across intervention groups was unclear in 10% of the studies. In summary, 45% of the NRS were considered at moderate, serious or critical risk of bias. Conclusions: A significant flaw in the assessment of adverse outcomes was found in a sample of NRS designed to investigate a pre-specified adverse effect. It is expected that a substantial error in the detection of harm outcomes may also be found in intervention trials. Further research on the impact of differential measurement error on the detection of adverse outcomes is required. RO 1.4 Inter-rater reliability of the Risk Of Bias Assessment Tool: for Non-Randomized Studies of Interventions (ACROBAT-NRSI) Couto E1 , Pike E1 , Torkilseng EB1 , Klemp M1 1 Norwegian Knowledge Centre for the Health Services, Norway Background: We conducted a systematic review to examine the possible association between the use of benzodiazepine derivatives and benzodiazepine related drugs and mortality. Non-randomised studies were included in the systematic review. We assessed risk of bias using the Cochrane Risk Of Bias Assessment Tool: for Non-Randomized Studies of Interventions (ACROBAT-NRSI). Objectives: To present the agreement between reviewers who used ACROBAT-NRSI to assess the risk of bias of studies included in this systematic review. Methods: Three reviewers (with experience in assessing risk of bias) independently examined the risk of bias of 33 prospective cohort studies. Inter-rater reliability of the bias assessment was estimated by calculating kappa statistics (k) using Stata. This was performed for each domain of bias separately and for the final overall assessment. Agreement was categorized as poor (K < 0.01), slight (k = 0.01 to 0.20), fair (K = 0.21 to 0.40), moderate (K = 0.41 to 0.60), substantial (K = 0.61 to 0.80), or almost perfect (0.81 to 1.00). Results: Overall, assessed studies were judged as being at serious or critical risk of bias. There was slight agreement for the overall assessment of risk of bias (k = 0.03). For the seven domains of bias, agreement ranged from k = -0.19 to K = 0.35. There was poor agreement for two domains (bias in measurement of interventions and due to departures from intended interventions), slight for two (bias in selection of participants into the study and in selection of the reported result), and fair for three (bias due to confounding, due to missing data, in measurement of outcomes). Conclusions: Cochrane Database Syst Rev Suppl 1–327 (2015) 10 The inter-rater agreement varied across the domains of the risk of bias tool, ranging from poor to fair. While we had slight agreement for the overall assessment of risk of bias, all reviewers independently assessed overall risk of bias of examined studies to be either serious or critical. included trials of adequate quality based on a ≥ 6 cut-off. Nineteen meta-analyses did not include trials of adequate quality according to the Cochrane approach. Agreement between PeDro and Cochrane was poor for PeDro scores of ≥ 5 points (k = 0.12; 95% CI 0.07 to 0.16), slight agreement for ≥ 6 points (k = 0.24; 95% CI 0.16 to 0.32), and ≥ 7 points (k = 0.39; 95% CI 0.286 to 0.510), and fair agreement (k = 0.44; 95% CI 0.314 to 0.574) ≥ 8 points (Figure 1). Conclusions: The PeDro and Cochrane approaches lead to different sets of trials of adequate quality. A consistent approach to assessing RoB in trials of PT should be adopted. RO 1.6 Developing and validating a tool to quantify risk of bias in systematic reviews of genetic association studies Sohani Z1 1 McMaster University, Canada RO 1.5 PEDro or Cochrane to assess the quality of physiotherapy trials? A meta-epidemiological study Armijo-Olivo S1 , da Costa BR2 , Cummings GG1 , Ha C1 , Fuentes J1 , Saltaji H1 , Egger M3 1 University of Alberta, Canada; 2 Institute of Primary Health Care, University of Bern, Switzeralnd; 3 Institute of Social & Preventive Medicine (ISPM), University of Bern, Switzerland Background: The use of different tools for evaluating the quality of randomized controlled trials (RCTs) in systematic reviews (SRs) can lead to discrepancies and skewed interpretations of SR results. The PeDro tool, which uses summary scores to determine the quality of RCTs is commonly used in physical therapy (PT). An alternative approach is based on an assessment of individual components such as the items used in the Cochrane ’Risk of bias’ (RoB) tool. Objective: To determine the agreement between the Cochrane and the PeDro approaches to identifying physiotherapy trials of adequate quality. Methods: We conducted a meta-epidemiological study. RCTs in PT were identified by searching the Cochrane Database of Systematic Reviews for meta-analysis of PT interventions. Final scores for PeDro and Cochrane assessments were extracted. We defined trials of adequate quality as having adequate sequence generation, concealment of allocation, and blinding of outcome assessors (based on the Cochrane RoB tool) or as trials with a PEDro score of at least 5 or 6 points, the cut-offs widely used in the literature. We calculated the kappa statistics to assess the agreement between PeDro scores and the Cochrane approach for classifying trial quality. We compared effect sizes from trials of adequate quality for each meta-analysis between the two approaches. Results: Forty-one Cochrane reviews and 353 trials of PT interventions were included. All meta-analyses included trials of adequate quality based on a ≥ 5 PEDro points cut-off and 37 (87.8%) Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Background: Completion of the human genome project and advances in genotyping technology has resulted in an increase in publication of genetic association studies. Systematic reviews and meta-analyses are a common method of providing synthesized effect of a genetic variant on a trait of interest, but, summary estimates are subject to bias due to the varying methodological quality of individual studies. Objective: To quantify bias introduced from individual studies and inform meta-analyses, we developed, validated, and empirically evaluated a tool that assesses the quality of genetic association studies. Methods: We used published guidelines and recommendations to create a list of items with potential impact on quality. We chose final items in consultation with five experts. Evaluation of the tool was performed in two parts. Firstly, four reviewers rated 30 studies randomly selected from a published meta-analysis. Their ratings were used to assess construct validity, reliability, and item discrimination of the tool. We report G-coefficients as measures of inter-rater reliability, internal consistency, and overall reliability of the tool, as well as item-total correlations and Cronbach’s alpha to assess the discriminative ability of each item. Secondly, a systematic review of published meta-analyses was conducted. The tool was applied to 50 randomly selected meta-analyses from the literature and studies of ‘poor’ quality were excluded to assess whether application of our tool improves precision of estimates and reduces heterogeneity. Results: The tool demonstrates excellent psychometric properties and generates a quality score for each study with corresponding ratings of ‘low’, ‘moderate’, or ‘high’ quality. When applied to individual meta-analyses to exclude studies of low quality, we found a decrease in heterogeneity and an increase in precision of summary estimates. Conclusion: Integration of our tool into meta-analyses to inform selection of studies for inclusion, conduct sensitivity analyses, and perform meta-regressions, can help improve the state of evidence in the field of genetic epidemiology, which is currently plagued with irreproducible findings. Cochrane Database Syst Rev Suppl 1–327 (2015) 11 RO 1.7 RO 1.8 PROBAST: a risk of bias tool for prediction modelling studies AMSTAR: helping decision makers distinguish high and low quality systematic reviews that include non-randomized studies Wolff R1 , Whiting P2 , Mallett S3 , Riley R4 , Westwood M1 , Kleijnen J5 , Moons K6 1 Kleijnen Systematic Reviews Ltd, United Kingdom; 2 University of Bristol, United Kingdom; 3 University of Birmingham, United Kingdom; 4 University of Keele, United Kingdom; 5 Kleijnen Systematic Reviews Ltd/ Maastricht University, United Kingdom/ The Netherlands; 6 University of Utrecht, The Netherlands Background: Quality assessment of included studies is a crucial step in any systematic review (SR). Review and synthesis of prediction modelling studies is a relatively new and evolving area and a tool facilitating quality assessment for prognostic and diagnostic prediction modelling studies is needed. Objectives: To introduce PROBAST (prediction study risk of bias assessment tool), a tool for assessing the risk of bias (RoB) and applicability of prediction modelling studies in a SR. Methods: A Delphi process, involving 42 experts in the field of prediction research, was used until agreement on the content of the final tool was reached. Existing initiatives in the field of prediction research such as the REMARK (reporting recommendations for tumour marker prognostic studies) and TRIPOD (transparent reporting of a multivariable prediction model for individual prognosis or diagnosis) reporting guidelines formed part of the evidence base for the tool development. The scope of PROBAST was determined with consideration of existing tools, such as QUIPS (quality in prognostic studies) and QUADAS (Quality assessment of diagnostic accuracy studies). Results: After 6 rounds of the Delphi procedure, a final tool was developed that utilises a domain-based structure supported by signalling questions similar to QUADAS-2. PROBAST assesses the RoB and applicability of prediction modelling studies. RoB refers to the likelihood that a prediction model leads to distorted predictive performance for its intended use and targeted individuals. The predictive performance is typically evaluated using calibration, discrimination, and (re)classification. Applicability refers to the extent to which the prediction model from the primary study matches the SR question, for example in terms of the population or outcomes of interest. PROBAST comprises 5 domains (participant selection, outcome, predictors, sample size and flow, and analysis) and 22 signalling questions grouped within the domains. Conclusions: PROBAST can be used to assess the quality of prediction modelling studies included in a SR. The presentation will give an overview of the development process and introduce the final tool. Shea B1 , The AMSTAR Development Group TADG2 1 OHRI, Ottawa Hospital, University of Ottawa, Canada; 2 Bruyere Research Institute, Bruyere Continuing Care, Canada Background: AMSTAR (A Measurement Tool to Assess Systematic Reviews) is widely used in the critical appraisal of systematic reviews (SRs). The instrument was validated using SR of randomized controlled trials (RCTs). Surveys have shown that over 60% of SRs of treatment effects include non-randomized studies (NRS). Objectives and methods: To develop a version of AMSTAR capable of appraising SRs that include NRS. A literature review found no critical appraisal instrument that incorporated suitable measures of risk of bias (RoB) of the component studies. An expert group reviewed the results and suggested a single instrument with separate paths to incorporate RoB assessment of either RCTs or NRS. The opportunity was taken to update the AMSTAR instrument in a number of other domains. Results: The modified AMSTAR instrument now has 15 items. The review process is guided more clearly by the PICOT framework (i.e. population, intervention, comparison, outcome, and timeframe), makes a more detailed assessment of literature searching methods, gives guidance on whether authors have made adequate assessment of RoB of included studies and whether limitations were acknowledged in making conclusions from the results of the review. The domains of bias (modified from those in the ACROBAT-NRSI tool) are: freedom from residual or unmeasured confounding; freedom from selection bias; freedom from bias in measuring exposures and outcomes; freedom from selective outcome/analysis reporting. The revised instrument is undergoing pilot testing. Conclusions: assessing the quality of SRs that include NRS is very challenging. The modified AMSTAR is a first step towards a more accurate appraisal of this important type of study. Rapid oral session 2 Editorial issues RO 2.1 What evidence is available and what is required, in humanitarian assistance? Allen C1 , Clarke M1 , Puri J2 , Archer F3 , Wong D3 , Eriksson A4 1 Evidence Aid, United Kingdom; 2 3ie, India; 3 Monash University, Australia; 4 Karolinska Institutet, Sweden Background: Challenges of conducting impact evaluations of interventions, actions and strategies in natural disasters and complex political emergencies are well recognized. Without a rigorous scoping study, there is no single resource informing researchers what the priority areas for impact evaluations (IEs) Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Cochrane Database Syst Rev Suppl 1–327 (2015) 12 are, nor a mapping of the gaps that exist in current research. Objectives: To investigate the current landscape of evidence, with emphasis on evidence from IEs, in the humanitarian sector to identify areas in which actionable evidence is available and where evidence is needed to direct research. Methods: The study incorporated multiple methods to assess available evidence including an online survey of participants knowledgeable in the humanitarian sector, semi-structured interviews with experts in the sector, extensive literature reviews of repositories of humanitarian studies and strategy documents of major humanitarian organisations and a gap map presenting results of a thorough search for completed, ongoing and planned IEs of humanitarian interventions. Results: With the exception of health and nutrition, most areas in the humanitarian sector suffer from a paucity of evidence. An evidence gap map provides an illustration of the landscape of evidence. There is agreement amongst policymakers that decisions should be based on research evidence and that IEs can and should have a greater role to play in building the evidence base. Conclusions: Key recommendations of this study, beyond informing areas and questions for IEs are that: Humanitarians must agree upon ways of prioritising research needs. This study suggests a framework for prioritizing further research. - Efforts should be made to index and classify existing evidence and a single unifying repository or portal would to improve the ease of accessibility to existing evidence. - A single set of templates and reporting guidelines should be agreed upon to aid in the indexing and classification of evaluation studies. It would also be beneficial to agree upon standards for data collection in these studies. RO 2.2 Evidence Aid Resources: improving access to systematic reviews that are relevant to disasters Aburrow T1 , Allen C2 , Clarke M2 1 Wiley, United Kingdom; 2 Evidence Aid, United Kingdom Background: Evidence Aid, with partners, assesses published systematic reviews (SRs) to determine their relevance to natural disasters, humanitarian crises or major healthcare emergencies. The SRs will have health outcomes and are featured on www.EvidenceAid.org/. Resources with a summary of each review and link to the full text. The main source of SRs for Evidence Aid continues to be Cochrane Reviews, but several non-Cochrane SRs have recently been added. Most of the Cochrane Reviews are also included in the relevant Evidence Aid Special Collection (www.thecochranelibrary.com). Objectives: To provide a curated and unique collection of SRs to facilitate an informed response by those preparing for and responding to natural disasters and humanitarian crises. Methods: Evidence Aid works closely with Cochrane and other external partners to identify evidence with health outcomes that decision-makers can assimilate quickly. Members of the Evidence Aid team (largely volunteers), provide a brief summary of the SR, which Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. is featured alongside a link to the full article. Wiley’s editorial team provides support to publish the resources, and relevant information, on a dedicated website: www.EvidenceAid.org. As SRs are added, they are communicated across social media channels. Cochrane and Wiley have provided free one-click access to the full version of all Cochrane Reviews featured in the Evidence Aid resource, to everyone in the world and, recently, Elsevier agreed to provide free click-through access (from the Evidence Aid website) to a review published in one of their journals. Results: Since November 2011, over 200 titles have been published in the resources collection and these titles are split into 39 categories. The /resources page is the most viewed on the site after the homepage, and on average holds a 12% share of all traffic. Conclusions: Evidence Aid provides a single source of health-related SRs and specially prepared summaries that might be relevant to those preparing for and responding to natural disasters and humanitarian crises. Evidence Aid continues to identify SRs and works with publishers to secure free, global access to the reviews. RO 2.3 Reviews in the humanitarian field: adaptation and transparency Krystalli R1 , Ott E2 , Stites E1 1 Feinstein International Center, Tufts University, USA; 2 Oxfam GB, United Kingdom Issue: Humanitarian actors and researchers have amassed evidence, but synthesizing this information and communicating it to policymakers and practitioners remains challenging. Although all reviews face methodological challenges, the context for humanitarian evidence is fundamentally different from the medical trial context from which systematic reviews emerged. The settings –which often involve active armed conflict, natural disasters, or the immediate aftermath of conflict or disasters –render data collection difficult or impossible. When data exist, the reliability, representativeness, and generalisability may be limited by access and numerous biases, including selection, recall, and reporting bias. Even in humanitarian settings that are not characterized by instability or insecurity, the types of studies can be limited by the fragility of protracted crises, the sensitivity of the information, and issues related to identifying and accessing affected populations. Response: As a response to this context and the low-level of understanding around systematic reviews in the humanitarian field, the Humanitarian Evidence Programme brought together and adapted existing literature for a guidance note on evidence synthesis in the humanitarian sector. Building on work from Cochrane, EvidenceAid, Overseas Development Institute, and 3ie, this note presents technical guidance in a manner accessible to an audience with a strong humanitarian identity. One challenge is to synthesize and communicate ’low-confidence’ qualitative evidence. All of our evidence synthesis outputs strive to be transparent about which Cochrane Database Syst Rev Suppl 1–327 (2015) 13 evidence they synthesize, clear about gaps and limits in the literature, and systematic in their mode of analysis of existing research. The guidance note is currently being used for a number of reviews, from acute malnutrition to shelter and settlement strategies. As with other reviews, the Humanitarian Evidence Programme reviews hope to identify gaps in existing research and knowledge, form conclusions with greater confidence than with individual studies, showcase disagreement and diversity among the literature, and highlight opportunities for further research. RO 2.4 Evidence aid: improving the science and evidence base of disaster response - a policy Delphi engagement Allen C1 , Jillson I2 , Clarke M1 , Waller S3 , Koehlmoos T4 , Mellon D1 , Kayabu B1 1 Evidence Aid, United Kingdom; 2 Georgetown University, USA; 3 Uniformed Services University of the Health Sciences, USA; 4 US Marine Corps, USA Background: The Policy Delphi Technique is a policy analysis method used to engage groups of experts in dialogue regarding a particular issue of focus. The Policy Delphi approach entails a series of iterative, structured dialogues, with each one referred to as a ‘round’ and building on the one before. Respondents answer a series of questions anonymously, the responses are summarized and used as a basis for subsequent questions. Questions may be added as the questionnaires are developed. Policy Delphi has been used in disaster management since the 1970s; differing from the traditional Delphi approach, it does not seek consensus, but explores alternatives and their implications. Objectives: The proposed policy Delphi would be conducted between May and November 2015 as a means of preparing for the proposed 2015 Evidence Aid meeting in the USA in November. The purpose of the policy Delphi is to engage a wide range of key stakeholders in dialogue regarding disaster responses, including for example, evidence for best practices in disaster response, approaches to improving investments in disaster response with Cochrane-style (systematic review) analysis of evidence, identification of gaps in the evidence base for disaster response; and factors that impact on effective disaster response. Methods: Three rounds: the first and second online with around 100 selected panelists and the third being conducted as part of the meeting. The broad research questions are: To what extent is evidence for best practices in disaster response available to a wide range of stakeholders? To what extent is Cochrane-style (systematic review) analysis used to assess evidence for best practices in disaster response? What are the most effective approaches to improving the cost-effectiveness of investments in disaster response? How can the ethical, legal and social issues related to disaster response decision-making be most effectively addressed? What are the factors that impact on effective disaster response decision-making? Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Results & Conclusions: Each round has an interim report with findings. The findings and conclusions from these data will be shared at the Colloquium. RO 2.5 Capacity-sharing for improved filtering of information for decision-making in the African context Stewart R1 , Langer L1 , Zaranyika H1 , Choge I1 , Erasmus Y1 , Jordaan S1 , Mitchell J1 , Gerald L1 1 University of Johannesburg, South Africa Introduction: There is limited capacity in systematic reviewing and evidence-use in low- and middle-income countries, particularly across Africa. This is reflected in the Global Evidence Synthesis Initiative (GESI) in which Cochrane plays a leading role. There are however significant pockets of capacity. The Africa Evidence Network, which we coordinate, seeks to build capacity by supporting existing initiatives that aim to filter information for better decisions in health care and other fields, amongst researchers, governments and non-governmental organisations (NGOs). This presentation will describe our activities, approaches and our preliminary evaluation, and invite the audience to share their own experiences. Background: The Network was formed in 2012 in an attempt to bring together these disparate initiatives and provide a forum for collaboration and capacity-development. Since 2014 the Network has been supported by the UK Department for International Development via our University of Johannesburg’s ‘Building Capacity to Use Research Evidence’ initiative. This BCURE programme supports the Africa Evidence Network website, regular newsletters and biannual Colloquia. We also run a capacity-building programme to increase use of research by the governments of Malawi and South Africa. Key Approaches: 1. A southern-based initiative, consciously avoiding ‘parachuting in’, instead working to ensure we build lasting relationships with the communities in which we work. 2. Using multiple media to exchange knowledge, including face-to-face interactions, social networks, podcasts etc., thus optimizing access to knowledge and networks beyond sectoral expertise. 3. Taking great care to ensure we build on existing initiatives, capacities and resources, to ensure we meet real needs and truly contribute to current capacity in research-use rather than merely replicating, duplicating or contradicting current efforts. 4. Avoiding the practice of training the trainers and instead adopting an apprenticeship approach. 5. Playing the role of matchmakers, fostering meaningful networks of research-users, research-producers and intermediaries in the countries in which we work. Cochrane Database Syst Rev Suppl 1–327 (2015) 14 RO 2.6 RO 2.7 Policy BUDDIES: lessons learnt by researchers engaging in ’buddying’ with provincial policymakers in South Africa to advance evidence-informed decision-making Story of an epidemic - public health evidence South Asia - creating a social epidemic by experimenting with few 1 2 1 3 4 Young T , Naude CE , Wiysonge CS , Kredo T , Dudley L , Garner P5 1 Centre for Evidence-Based Health Care, Stellenbosch University; South African Cochrane Centre, South African Medical Research Council, South Africa; 2 Centre for Evidence-Based Health Care, Stellenbosch University, South Africa; 3 South African Cochrane Centre, South African Medical Research Council, South Africa; 4 Division of Community Health, Stellenbosch University, South Africa; 5 Effective Health Care Research Consortium, Liverpool School of Tropical Medicine, United Kingdom Background: Successful dialogue between researchers and policymakers may facilitate evidence-informed decision-making by helping policymakers understand what research can or cannot do, and researchers understand how policy making works. Objectives: To develop and apply a formal one-to-one partnership (‘buddying’) between provincial policymakers and evidence-based health researchers (‘buddies’); and to evaluate the buddies’ experiences. Methods: Policymakers identified questions relevant to them during interviews in the baseline situational analysis. Each buddy initiated and drove the interaction process with a policymaker for their question, drawing support from the buddies group that included all researchers involved in the project. A dedicated online website and monthly contact sessions were used by buddies to share resources and continuously document interactions, reflections and experiences. Results: We applied buddying to five questions: (1) adherence support for chronic diseases; (2) prevention of mother-to-child transmission of HIV; (3) task shifting for antiretroviral treatment; (4) health promotion for diet and physical activity’ and (5) integration of care for chronic diseases. Having policymakers who champion the use of evidence appears to facilitate ‘buddying’. Engagement helped researchers understand that systematic review evidence is only a piece of the complex policymaking puzzle. Buddying can be time-consuming and is affected by frequent staff turnover of policymakers. Scenarios for evidence requests were different: for some questions, evidence was sought to endorse existing policies and for others, to inform new policy development. Conclusions: Successful, sustainable engagement between researchers and policymakers to advance evidence-informed decision-making is not a quick, simple process. Buddying, centred on partnerships and relationships, may provide a knowledge translation platform that could strengthen health interventions and services by enhancing evidence into policy; however, many uncertainties still remain. Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Nair N S1 , Mujja A2 , Lewis M3 , Nagaraja R3 , T Venkatesh B4 1 Director, Public Health Evidence South Asia, Manipal University, India; 2 Research Assistant, Public Health Evidence South Asia, Manipal University, India; 3 Research Scholar, Public Health Evidence South Asia, Manipal University, India; 4 Research Officer, Public Health Evidence South Asia, Manipal University, India Background: Public Health Evidence South Asia (PHESA) is an initiative aiming to meet the public health evidence needs of the South Asian region. The initiative, which includes the South Asian satellite of the Cochrane Public Health Group (CPHG), is based at Manipal University, India. Objectives: This paper describes the formation and successful journey of PHESA. This is a model that demonstrates that a few dedicated and effective people can create a social epidemic like PHESA - the theory of few. The paper talks on how infection began and to what extent it has spread over a period of two years from 2013. Methods: The first part of the paper concentrates on genesis of PHESA; the struggles involved in convincing Manipal University to support the activity, working out vision and mission, acquiring funding, logistics and identifying a dedicated workforce and so on. It also elucidates the activities of networking with organizations and individuals, sensitization workshops, conference presentations etc. The second part of the paper shares the success story of PHESA. Results: During the last two years PHESA has trained 150 researchers/students, collaborated with six national and six international organizations like Public Health Foundation of India (PHFI), the World Health Organization (WHO), Nossal Institute of Global Health, University of Melbourne and University of Groningen. Apart from this, PHESA has three full time PhD scholars registered with funding, 15 completed student internship projects and has attracted research funding of INR 8.5 million. The research output includes nine published papers, nine communicated papers and six papers in preparation. The first PHESA colloquium was organized in February 2015 with 75 participants, four plenary talks, 38 oral presentations, 25 posters and three workshops. Conclusions: The message we are experimenting is ’money is not all that matters but few contagious, sticky people- connectors and experts (mavens) - can make a big difference at the right context for public health in low- and middle-income countries’. The journey is still on. Cochrane Database Syst Rev Suppl 1–327 (2015) 15 RO 2.8 Use of Cochrane Reviews in developing Clinical Practice Guidelines (CPGs) in Malaysia Abdullah Thani NSI1 , Muhamad NA1 , Lai NM2 , Tan ML3 , Ho J3 1 Institute for Medical Research, Malaysia; 2 Taylor’s University, Malaysia; 3 Penang Medical College, Malaysia Background: Clinical Practice Guidelines (CPGs) are developed to improve the quality in healthcare practice based on the best available evidence. The process of developing the CPGs include systematically screening available evidence from systematic reviews. To-date, there have been 70 CPGs developed in Malaysia since 2004. Cochrane Reviews are widely regarded as the chief source of best evidence in health care that should be cited in the CPGs. Objectives: We examined the citation of Cochrane Reviews in Malaysian CPGs. Methods: Seventy CPGs were retrieved from the official Ministry of Health (MOH) Malaysia website: http://www.moh.gov.my/. The reference list of each CPG was screened manually to assess whether Cochrane Reviews were included. Two authors determined whether there was a key Cochrane Review that should be cited in each topic of the CPG without knowledge of the CPG references, and a third author assessed whether these key reviews were included in the corresponding CPG. Key reviews are large reviews with clear results and recommendations for practice. Results: Seventy CPGs covered 17 topics. A total of 219 (2.94%) Cochrane Reviews and 527 (5%) non-Cochrane reviews were Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. identified from 9893 references. A median of one Cochrane and four non-Cochrane reviews were cited per CPG (Table 1). From 2004 to 2014 Cochrane Reviews have been increasingly cited (Table 2). Out of 50 key Cochrane Reviews relevant to the CPG topics, 16 key reviews (30.8%) were included while 34 were not. Conclusions: Despite an encouraging trend of increasing citation, which suggests rising awareness of the importance of Cochrane Reviews, the inclusion of certain key Cochrane Reviews in the Malaysian CPGs remains unsatisfactory. RO 2.9 Measuring global health inequalities: the Gini, Theil and Slope inequality index for 291 diseases and injuries and 76 risk factors von Philipsborn P1 , Steinbeis F2 , Gotham D3 , Bender ME4 , Matthiessen C5 , Sauter C6 , Tinnemann P7 1 Technische Universät München and Universities Allied for Essential Medicines (UAEM), Germany; 2 Charité Universitätsmedizin Berlin and Universities Allied for Essential Medicines (UAEM), Germany; 3 Imperial College London, and Universities Allied for Essential Medicines (UAEM), United Kingdom; 4 Charité Universitätsmedizin Berlin, and Universities Allied for Essential Medicines (UAEM), Germany; 5 University of Copenhagen, and Universities Allied for Essential Medicines (UAEM), Denmark; 6 Universität Bonn, Germany; 7 Charité Universitätsmedizin Berlin, Germany Background: {Numerous health care initiatives, including Cochrane, aim to reduce global health inequalities. Knowledge about the contribution of different diseases and risk factors and existing disparities in health status can contribute to evidence-based, equity-sensitive priority setting in research and policy-making. Objectives: To calculate concise, comprehensive inequality measures to synthesize the Cochrane Database Syst Rev Suppl 1–327 (2015) 16 vast amount of information on health inequalities provided by the Global Burden of Disease Study 2010. Methods: We use the population-weighted Gini and Theil index to measure relative health inequalities, and the population-weighted Slope Inequality Index (SII) to measure absolute inequality, applying them to country-level disease burden data, measured in disability-adjusted life years (DALYs). The Gini index and the SSI are derived from the Lorenz Curve and the Pen’s Parade, respectively (see Graphs 1 and 2), which can also be used to visualize results. All three indices have been used widely in analysis of inequality in the distribution of health outcomes. Results: Overall relative and absolute global health inequalities increased between 1990 and 2000, but have decreased after 2000. Between 2000 and 2010, absolute inequality fell below 1990 levels, while relative inequality did not (see Graph 3). Cause groups that drove the rise in global inequality between 1990 and 2000 were HIV, interpersonal violence, and road traffic injuries. Overall, communicable, maternal, neonatal and nutritional disorders (Gini = 0.583), as well as injuries (Gin i= 0.302) contribute more to existing inequalities than non-communicable diseases (Gini = 0.172) (see Graph 4), with considerable variance within these groups. In 2010, relative global health inequalities among women (Gini = 0.318) were considerably larger than among men (Gini = 0.285). Conclusions: While global health inequality has decreased since 2000, overall levels remain high, with a marked variance amongst different health conditions and risk factors. Increased attention for those conditions and risk factors that contribute most to existing global health inequalities may be warranted. Rapid oral session 3 Advanced methods: Overviews of reviews and network meta-analyses RO 3.1 What guidance is available for researchers conducting overviews of reviews? a scoping review and qualitative metasummary Foisy M1 , Fernandes RM2 , Hartling L1 1 Alberta Research Centre for Health Evidence, University of Alberta, Canada; 2 Cochrane Child Health Field; Cochrane Portugal; Department of Pediatrics and Clinical Pharmacology Unit, Lisbon Academic Medical Centre, Portugal Background: Overviews of reviews (overviews) compile data from multiple systematic reviews to provide a single synthesis of relevant evidence for clinical and policy decision-making. Despite their increasing popularity, there is limited methodological guidance available for researchers wishing to conduct overviews. Objectives: To identify and collate all published and unpublished documents containing methodological guidance for the conduct or reporting of overviews. Our aim was to provide a map of existing guidance documents and identify similarities, differences, and gaps Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Cochrane Database Syst Rev Suppl 1–327 (2015) 17 in guidance. Methods: We conducted a database search supplemented by iterative snowball searching, handsearched conference proceedings and relevant websites, and contacted overview authors and Cochrane Review Groups. Guidance statements across included documents were grouped by stage of the overview process and analyzed within each stage. Results: We identified 48 guidance documents produced by 20 author groups: 29 contained explicit methodological guidance; 10 were peer-review checklists or reporting guidelines; and nine described the challenges an author group faced when conducting one or more published overviews. Overall, 79% of these documents were written by authors affiliated with Cochrane. General guidance was available for the following steps: deciding whether to conduct an overview, specifying the scope, searching, screening, and inclusion. This guidance was mostly consistent, though there were some minor discrepancies across included documents. There was limited or conflicting guidance on extraction and analysis of results data, methodological quality assessment, and grading the quality of evidence. Most documents identified additional challenges in conducting overviews for which they did not provide guidance. Conclusions: Compiling existing methodological guidance for overviews provides a valuable resource for overview authors. The results of this project will inform our research on overview methods and are being used to update the overviews chapter for the next edition of the Cochrane Handbook. RO 3.2 Scoping reviews versus systematic reviews: results from a scoping review of scoping reviews Tricco A1 , Lillie E1 , Zarin W1 , O’Brien K2 , Colquhoun H2 , Kastner M1 , Levac D3 , Ng C1 , Pearson Sharpe J1 , Wilson K1 , Kenny M1 , Warren R1 , Wilson C1 , Stelfox H4 , Moher D3 , McGowan J3 1 Li Ka Shing Knowledge Institute, St. Michael’s Hospital, Canada; 2 University of Toronto, Canada; 3 University of Ottawa, Canada; 4 University of Calgary, Canada Background: Scoping reviews are used to map concepts underpinning a research area and the main sources and types of evidence available. Currently, a lack of consistency exists in the terminology, definition, methods, and reporting of scoping reviews appearing in the literature. Objectives: We aimed to synthesize scoping review approaches through a scoping review. Methods: We conducted a scoping review of scoping reviews. Nine multi-disciplinary electronic databases were searched for scoping reviews or studies that discussed scoping review methodology (e.g. the Cochrane Library, PsycINFO, Philosopher’s Index). The citations were screened independently by two reviewers for inclusion, and data extraction was verified by a second reviewer. Descriptive analysis was conducted. Results: We included 545 scoping review articles. Preliminary results suggest that there has been an exponential increase in the number of scoping reviews conducted in the past five years. The majority Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. were conducted in North America (56%) and Europe (39%). Most studies reported a source of funding (61%) that was either from the government (25%) or a health authority (16%). The most commonly reported approaches to the conduct of scoping reviews included searching one or more database (92%), scanning reference lists (59%), searching grey literature (52%), limiting inclusion by date (68%) or language (62%), and not conducting a formal quality (risk of bias) appraisal (81%). In most cases, the results of the scoping review were used to provide recommendations for future research (78%), identify evidence gaps (76%), and/or to specify policy or practice recommendations (54%). Conclusions: The number of scoping reviews conducted per year is increasing exponentially. Scoping reviews are used to set research agendas and provide policy or practice recommendations, yet there is variability in their reporting and the methods that are being used. Further research on scoping reviews is warranted, in particular, there is need for a guideline to standardize reporting. RO 3.3 Overview of screening recommendations from evidence-based guidelines as decision support for a reoriented parent-child preventive care programme Reinsperger I1 , Piso B1 , Winkler R1 1 Ludwig Boltzmann Institute for Health Technology Assessment, Austria Background: Against the backdrop of the observed rapid increase in published clinical practice guidelines (CPGs), systematic CPG analyses (‘overviews of CPGs’) can be considered as information filters by summarising recommendations of evidence-based CPGs. Objectives: We aimed to provide the Ministry of Health with an overview of antenatal, perinatal and postnatal screening recommendations from international evidence-based CPGs as decision support for a reoriented national parent-child preventive care programme. Methods: A systematic search in two guideline databases was supplemented by a comprehensive handsearch via websites of international institutions developing CPGs. We included evidence-based guidelines from western industrialised countries providing screening recommendations for pregnancy and early childhood that were published or updated within the past five years. Guideline development had to be based on a systematic literature search and recommendations had to be explicitly linked to the evidence. We extracted the screening recommendations, grades of recommendation, the screening times and methods, treatment options and potential harms. Results: The systematic CPG analysis provides a comprehensive, systematic overview of recommendations from evidence-based guidelines of international institutions for screening measures during pregnancy and early childhood: We included a total of 138 guidelines, published by 14 institutions, and extracted Cochrane Database Syst Rev Suppl 1–327 (2015) 18 recommendations related to 92 health threats. Basically, grades of recommendations were high for all pro-screening recommendations, whereas certainty decreased across all contra-screening recommendations. Conclusions: The policy decision on the implementation of new screening measures or the adaptation of existing screening approaches requires the appraisal of the provided CPG analysis. Currently, an interdisciplinary expert panel discusses the transferability and applicability of the international CPGs and develops recommendations for the national context. development. Our Overview comprehensively, rigorously and effectively signposts clinicians to reviews that are most likely to support delivery of effective care. RO 3.5 Constructing an overview of systematic reviews of diagnostic test accuracy Hunt H1 , Hyde C1 1 University of Exeter Medical School, United Kingdom RO 3.4 Completing the first Cochrane overview of stroke reviews: experiences of the Cochrane Stroke Group Pollock A1 , Farmer S1 , Brady M1 , Langhorne P2 , Mead G3 , Mehrholz J4 , van Wijck F1 1 Glasgow Caledonian University, United Kingdom; 2 University of Glasgow, United Kingdom; 3 University of Edinburgh, United Kingdom; 4 Klinik Bavaria in Kreischa GmbH, Germany Background: Cochrane Overviews aim to provide a succinct overview of reviews relevant to a specific clinical question. Objectives: To explore the key methods associated with producing the first Cochrane Stroke Group Overview. Methods: Six editors from the Cochrane Stroke Group systematically explored methodological issues arising during the overview process. Monthly teleconferences and one full-day meeting were held to discuss and agree solutions to these issues. Each methodological issue, and how it was addressed, was recorded. Results: Our completed overview included 40 reviews (19 Cochrane, 21 non-Cochrane). Key methodological challenges encountered related to: 1. inclusion of non-Cochrane reviews: this added substantial complexity to review inclusion, primarily due to overlap between reviews, requiring systematic exploration of the inclusion of more than 700 studies; 37 reviews were excluded as they were superseded by more up-to-date or comprehensive reviews; 2. quality assessment of included reviews: we found use of the AMSTAR quality assessment tool challenging due to its use of many multifaceted questions. We developed a modified version of the AMSTAR tool (mAMSTAR), comprising 30 simple uni-variable questions. This enabled transparent documentation of the process of quality assessment. 3. quality of evidence within reviews: we planned to use the GRADE approach to assign levels of evidence to each of the 127 relevant comparisons included in the overview, but found it difficult to achieve consistent application. We therefore developed and used an objective algorithm based on participant numbers, trial risk of bias, heterogeneity and review quality. 4. strength of implications: only 1/127 comparisons was graded as high quality, and 76/127 comparisons were graded as low or very-low quality evidence, limiting clinical implications. Conclusions: Overviews should benefit evidence-based stroke care, but tools used to support Cochrane Overviews need further methodological Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Objectives: The aim of this presentation is to present an early career researcher’s experience of conducting an overview of the existing systematic review evidence for the diagnostic test accuracy of brief cognitive assessments used to identify dementia in primary care. Description: systematic reviews of existing health evidence such as those conducted by Cochrane are recognised as the international ’gold standard’ for high quality trusted information. As more systematic reviews are produced, there is a growing need to summarise and appraise the systematic review evidence in order to filter the information overload and give decision makers the best information available. However, whilst guidance on conducting an ‘Overviews of reviews’ is covered in Chapter 22 of the Cochrane Handbook for Systematic Reviews of Interventions (version 5.1.0), there is no published Cochrane guidance for authors conducting overviews of reviews of diagnostic test accuracy. In this presentation, the author will discuss: - What is the purpose of conducting an overview of systematic reviews? - How do Cochrane and non-Cochrane reviews differ, and should they be treated differently? What parallels and contrasts exist between accuracy and non-accuracy reviews? - How should an author handle a change in review perspective (e.g. primary care focus when reviewing general setting reviews)? - What would a good overview of systematic reviews of diagnostic test accuracy look like? The author aims to situate the evidence for overviews of systematic reviews of diagnostic test accuracy, and prompt discussion of issues raised. RO 3.6 Including observational evidence in a network meta-analysis of RCTs: a case study in schizophrenia Efthimiou O1 , Mavridis D2 , Debray T3 , Leucht S4 , Samara M5 , Salanti G1 1 Department of Hygiene and Epidemiology, University of Ioannina School of Medicine, Greece; 2 Department of Primary Education, University of Ioannina, Greece; 3 The Dutch Cochrane Centre, Julius Center for Health Sciences and Primary Care, University Medical Center Utrecht, The Netherlands; 4 Department of Psychiatry and Psychotherapy, Technische Universität München, Klinikum rechts der Isar, Ismaningerstr, Germany; 5 Department of Psychiatry and Psychotherapy, Technische Universität München, Klinikum rechts der Isar, Ismaningerstr., Germany Cochrane Database Syst Rev Suppl 1–327 (2015) 19 Background: Observational studies convey valuable information about the effectiveness of interventions in real-life clinical practice. Recent research has found no evidence of systematic differences in the estimated effects between observational and randomized controlled trials (RCTs). However, there is little available guidance on how to combine real-world data with evidence from RCTs, especially when there are multiple competing available treatments for the same condition. Objectives: To assess existing methodology and develop new methods for jointly synthesizing evidence on relative treatment effects provided by RCTs as well as evidence from observational studies that report individual patient data (IPD) in a network meta-analysis (NMA). Methods: We adjusted the relative treatment effects from the observational studies utilizing patient level covariates. We combined the adjusted estimates with the evidence provided by RCTs in a joint network meta-regression, accounting for trial-level covariates, and exploring a variety of alternative methods. We quantified the impact of different levels of confidence for the adjusted estimates on the NMA relative effects. Results: We applied our methods in a published network of 167 RCTs, comparing 15 antipsychotics and placebo for schizophrenia. We extended our evidence base by including IPD from a cohort study involving five interventions and more than 10,000 patients. The adjusted evidence was found to be in agreement with the evidence from RCTs. Statistical inconsistency in the network was unaffected by the inclusion of observational evidence. A range of sensitivity analyses shows our results to be robust to the alternative methods used. Conclusions: Including real-world evidence from observational studies can corroborate findings of a NMA based on RCTs alone, increase precision and enhance the decision-making process. where we first used random-effects, found counter-intuitive results, and then applied fixed-effects models. Results: Applying the GRADE (Grading of Recommendations Assessment, Development and Evaluation) approach, we used random-effects models to generate direct, indirect, and network estimates. We found instances where direct and indirect estimates were similar, but CIs of NMA estimates were far wider than the direct estimates. In these instances, fixed-effects models generated narrower CIs consistent with the direct estimates, e.g. in an NMA of alternative fluids for resuscitation in septic shock, random-effects showed very similar point estimates for direct and indirect comparisons; the CI around the direct estimate was far narrower (Table). Counter-intuitively, the NMA estimate was far wider than the direct estimate. In the fixed-effect model, the sparse data from the indirect estimate changed the NMA estimate little in comparison to the direct estimate –as one would intuitively expect (Table). We have encountered a number of such situations. As in this example, in the face of similar point estimates from direct and indirect comparisons, basing clinical decision-making on an NMA estimate that is far less precise than the direct estimate is inappropriate. Options are to use the direct estimate only, or apply a fixed-effects model that generates sensible results. Conclusions: Those conducting Bayesian NMA need to be aware of potential problems with random-effect models and, when counter-intuitive results arise, consider fixed-effects models. RO 3.7 RO 4.1 Problems with Bayesian random effects in network meta-analysis The PRISMA-IPD Statement: preferred reporting items for a systematic review and meta-analysis of individual participant data Guyatt G1 , Murad H2 , Heels-Ansdell D1 , Puhan M3 1 McMaster University, Canada; 2 Mayo Clinic, USA; 3 Institute for Social and Preventative Medicine, Zurich, Switzerland Background: In most network meta-analyses (NMAs), substantial variability in results makes random-effects an appealing model. A potential advantage of NMAs is the possibility that consistent results from direct andindirect comparisons will narrow confidence intervals (CIs) in comparison to those from direct estimates alone. If direct comparisons provide higher quality evidence than network estimates, clinicians should use that higher certainty evidence to guide their practice. Objective: To illustrate problems in the use of random-effect models in Bayesian NMAs. Method: We conducted a number of Bayesian NMAs Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Rapid oral session 4 Quality of reporting Stewart L1 , Clarke M1 , Rovers M2 , Riley R1 , Simmonds M1 , Stewart G1 , Tierney J1 1 IPD Meta-analysis Methods Group, United Kingdom; 2 IPD Meta-analysis Methods Group, The Netherlands Background: Systematic reviews with meta-analysis of individual participant data (IPD) aim to collect, check, and re-analyze individual-level data from all studies addressing a particular research question and are considered a gold standard approach to synthesis. Like most areas of research, reporting could be improved, making it easier for readers to understand, critique and implement findings. Standard PRISMA guidelines are geared towards systematic reviews based on aggregate data and so lack reference to Cochrane Database Syst Rev Suppl 1–327 (2015) 20 some important aspects of the IPD approach. Objectives: The PRISMA-IPD extension was developed to provide a framework for full and transparent reporting of IPD review methods. Methods: Development followed the EQUATOR Network framework guidance and used the standard PRISMA statement as a starting point to draft additional relevant material. A web-based survey informed discussion at an international workshop. The statement was drafted and refined iteratively and agreement on the PRISMA-IPD checklist and flow diagram agreed by consensus. Results: PRISMA-IPD contains 23 items in which the wording has been modified to take reflect IPD approaches, and three new items on: (1) methods of checking data integrity; and (2) reporting any important issues identified; and (3) on methods of exploring variation and one new item from re-arrangement. Although developed primarily for reviews of randomized trials, many items will apply in other contexts including for reviews of diagnosis and prognosis. Conclusions: PRISMA IPD includes a tailored checklist and flow diagram, which we hope will improve reporting. If, as a result of current initiatives aiming to make provision of clinical trial data for research purposes a legal, regulatory or ethical requirement IPD become more readily available, then it is likely that in the future more systematic reviews will access and analyze IPD. This will likely include synthesis of IPD released in controversial areas where transparent, complete and high quality reporting is essential. Authors and peer reviewers are encouraged to use to improve reporting and journal editors to include it in their endorsement of PRISMA. RO 4.2 The Consensus on Exercise Reporting Template (CERT): a Delphi study investigating a standardised method for reporting exercise programs Slade S1 , Dionne C2 , Underwood M3 , Buchbinder R4 1 Monash University, Department of Epidemiology and Preventive Medicine, Melbourne, Australia; 2 Monash Department of Clinical Epidemiology, Cabrini Hospital, Melbourne, Australia; 3 Department of Rehabilitation, Faculty of Medicine, Laval University, Québec, Canada; 4 Warwick Clinical Trials Unit, Division of Health Sciences, Warwick Medical School, University of Warwick, Coventry, United Kingdom Background: Exercise is effective, integral to health and important for people with chronic health conditions. A systematic review of 73 reviews of exercise for chronic conditions reported suboptimal descriptions of interventions and concluded that this hinders replication and implementation. Without explicit descriptions, clinicians, patients and researchers remain unclear about effective programs. Objectives: The aim was to develop a standardised method for reporting key and essential exercise program details evaluated in clinical trials. Methods: A modified Delphi technique was used to gain consensus among an international panel of experts identified by systematic review authorship, Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. international research and clinical profile and peer referral. An online survey of 42 questions was developed from our systematic review. Experts from 14 countries indicated item importance on a scale of 0 to 10 and gave free text comments. We used three sequential rounds of anonymous online questionnaires. The final template will be piloted on a random sample of systematic reviews and RCTs of exercise. Modifications will be made and final approval sought from the expert panel. Results: There were 57/137 respondents in Round 1 (42%), 10 items were accepted in the original format and 14 reformatted for Round 2 which had 53/57 respondents (93%). Sixteen items were distributed for Round 3 which had 49/54 respondents (91%). The final core set was 16 items and 520 comments were thematically analysed. The final template was modeled according to the CONSORT Statement and TIDieR Checklist. An Explanation and Elaboration Statement and implementation strategies such as journal endorsement for manuscript submission and peer-review and CONSORT Item 5 extension are planned. The CERT checklist is registered on the Equator Network. Conclusions: The CERT checklist will improve explicit reporting of exercise interventions in clinical trials; increase clinical uptake of effective exercise programs; enable research replication; improve patient outcomes; and increase research impact on health. RO 4.3 Essential items in reporting diagnostic accuracy studies: 2015 update of the STARD Statement Bossuyt PM1 , Korevaar DA1 , Hooft L2 , Cohen JF3 , The STARD Group &4 1 Department of Clinical Epidemiology, Biostatistics and Bioinformatics, AMC, University of Amsterdam, The Netherlands; 2 Dutch Cochrane Centre, Julius Center for Health Sciences and Primary Care, University of Utrecht, The Netherlands; 3 INSERM UMR 1153, Obstetrical, Perinatal and Pediatric Epidemiology Research Team, Center for Epidemiology and Statistics Sorbonne Paris Cité, Paris Descartes University, France; 4 (85 members, International) Background: Many studies that evaluated the diagnostic accuracy of one or more medical tests, by comparing test results against a clinical reference standard, fail to report key elements of the design and the results. This makes judgments about risk of bias and applicability of study findings difficult, and complicates evidence synthesis. In 2003, a group of researchers, editors, and other stakeholders published the STARD statement (STAndards for Reporting Diagnostic accuracy): a list of 25 essential items that should be included in every report of a diagnostic accuracy study. Evaluations have revealed modest, but significant, improvements in reporting since then. Objectives: In 2013, the STARD steering committee decided to prepare an update. Two considerations were key: (1) the update should be based on the accumulated evidence about sources of bias and variability and honest reporting; (2) the update should make STARD easier to use. Methods: Original STARD group members were invited to participate, and to nominate new members. The STARD group now Cochrane Database Syst Rev Suppl 1–327 (2015) 21 consists of 85 researchers, editors, and other stakeholders. In web-based surveys, they could respond to and suggest changes in the list of items. A draft version of the new list of items was put together during a live meeting of the steering committee and then piloted. Results: The updated list consists of 30 items. Compared to the 2003 version, a few items were added, some were combined, others were split into separate items, while the remaining items were sometimes reworded, to make them easier to understand, and to harmonize wording with other reporting guidelines. Conclusion: The updated STARD list will be released later in 2015. As establishing a list of essential items is not sufficient to improve reporting, the release of the update will be accompanied by additional initiatives, such as explanatory documents, lectures, writing templates for authors and for reviewers. Improving the completeness and transparency of reporting will help us all to reduce avoidable waste in research, and strengthen the evidence base for making decisions and building recommendations about medical tests. RO 4.4 Quality of reporting of diagnostic accuracy studies in medical laboratory journals Pecoraro V1 , Banzi R2 , Trenti T1 1 Ospedale Civile Sant’Agostino Estense, Modena, Italy; 2 IRCCS-Istituto di Ricerche Farmacologiche Mario Negri, Milano, Italy Background: The Standards for Reporting of Diagnostic Accuracy (STARD) checklist aims to improve the reporting of diagnostic accuracy studies. So far, only a few journals have adopted the STARD checklist. Objective: We investigated the current quality of reporting of diagnostic accuracy studies focusing on laboratory medicine. Methods: We searched for journals formally adopting the STARD in the ’Medical laboratory technology’ category of ISI Web of Knowledge database. As a comparator, we included matching journals that do not adopt the checklist. We selected studies reporting an estimate of accuracy (sensitivity, specificity, area under ROC curve, likelihood ratio, predictive values, diagnostic odds ratio) published in 2014. For each article, we analyzed the compliance to STARD items and classified the reporting Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. as poor (less than 10 items reported), acceptable or optimal (more than 20 items reported). Results: Only three of 31 journals listed in the field of laboratory medicine mentioned the STARD checklist in their instructions for authors. We evaluated 24 studies published in these journals (Clinical Chemistry; Clinical Chemistry and Laboratory Medicine, and Annals of Clinical Biochemistry) and seven studies published in Chimica Clinica Acta which does not adopt the STARD (data collection is ongoing on other journals). Overall, the median number of items reported was 11 (range: 5 to 18). None of the studies complied with the entire checklist. Reporting was evaluated as poor in 15 of 31 studies (48.4% Table 1). The most frequently reported items were index test and reference standards. Papers often lacked an accurate description of the population enrolled. Conclusions: The majority of laboratory medicine journals do not adopt the STARD checklist. Some of the STARD items may be irrelevant to the studies being evaluated, however, preliminary data suggest that the use of the STARD checklist improves the quality of reporting. Final data on quality of reporting of journals adopting or not adopting the STARD checklist will be presented. Inadequate reporting of study participant characteristics may limit the relevance and application of results in practice. RO 4.5 Quality of reporting of clinical prediction model studies: adherence to TRIPOD Heus P1 , Damen J1 , Scholten R1 , Reitsma J1 , Collins G2 , Altman D2 , Moons K1 , Hooft L1 1 Dutch Cochrane Centre, Julius Center for Health Sciences and Primary Care, University Medical Center, Utrecht, The Netherlands; 2 Centre for Statistics in Medicine, NDORMS, Botnar Research Centre, University of Oxford, United Kingdom Background: There is a growing number of prediction models, both diagnostic and prognostic, that are published in the medical literature. Systematic reviews are required to deal with this information overload. However, systematic review authors are highly dependent on the quality of the reporting of primary studies. To improve the reporting of prediction models, a guideline for Transparent Reporting Cochrane Database Syst Rev Suppl 1–327 (2015) 22 of a multivariable prediction model for Individual Prognosis Or Diagnosis (TRIPOD) was launched in January 2015. The TRIPOD statement is a checklist of 22 main items considered essential for good reporting of studies developing or validating multivariable prediction models. Objectives: The objective of our study is to assess the quality of reporting of prediction model studies published before the launch of TRIPOD. In addition, this study could possibly serve as a baseline measurement for future studies evaluating the impact of the introduction of TRIPOD. Methods: For 37 clinical domains we selected 10 journals with the highest impact factors. A PubMed search was performed to identify prediction models published in May 2014. Publications that described the development and/or validation of a prediction model, either diagnostic or prognostic, were eligible. TRIPOD items were carefully translated into a data extraction form, which was piloted extensively. Consensus was reached on when to consider an item ’adhered’. Results: Our search identified 4871 references, 347 of which were potentially eligible references and were assessed in full text. Eventually 180 references were included. We will present the adherence to TRIPOD per publication, as well as per item, across studies, and across clinical domains. Also, diagnostic and prognostic models will be distinguished, and results will be separated for studies addressing model development, model validation, incremental value, or a combination of these. Conclusions: Our study will provide insight into the current quality of published reports about the development or validation of prediction models across a wide variety of clinical domains. Reporting issues will be identified that might require specific guidance. of bias tool, and the proliferation of new avenues to disseminate SRs. We considered it timely to evaluate a more recent cross-sectional sample. Objective: To investigate the epidemiological and reporting characteristics of SRs. Methods: We searched MEDLINE to identify potentially eligible SRs indexed in February 2014. Citations were screened using pre-specified eligibility criteria. We developed a standardised, pilot-tested data extraction form comprising all items used in [1], along with additional items to capture issues not previously examined. Example items include focus of the SR (e.g. therapeutic, diagnostic), number of included studies, and reporting of eligibility criteria, search strategy, and statistical analyses. Data in 300 SRs will be independently extracted by one of several reviewers, with a 10% random sample extracted in duplicate. We will compare characteristics of i) Cochrane versus non-Cochrane reviews, and ii) the current versus 2004 sample of SRs. Results: We identified 698 SRs indexed in February 2014. This suggests an annual publication rate of approximately 8300 SRs (equivalent to 23 SRs per day), which is more than triple the estimated rate in 2004. Cochrane Reviews comprised only 14% of the sample. Reporting characteristics will be presented at the Colloquium. Conclusion: This study will provide up-to-date data on the different types of SRs being published (e.g. therapeutic, diagnostic, prognostic) and the reporting quality of SRs. The findings will indicate where methodological and reporting guidance for SRs is most needed, and will inform the SR methodology research agenda. Reference: Moher D, Tetzlaff J, Tricco A C, Sampson M, Altman D G. Epidemiology and reporting characteristics of systematic reviews. PLoS Medicine 2007: 4(3): e78. RO 4.6 Epidemiology and reporting characteristics of systematic reviews: 2014 update Page MJ1 , Shamseer L2 , Altman DG3 , Tetzlaff J4 , Sampson M5 , Tricco AC6 , Catalá-López F7 , Li L8 , Reid E9 , Sarkis-Onofre R10 , Moher D2 1 School of Public Health and Preventive Medicine, Monash University, Australia; 2 Ottawa Hospital Research Institute and University of Ottawa, Canada; 3 University of Oxford, United Kingdom; 4 Ottawa Hospital Research Institute, Canada; 5 Children’s Hospital of Eastern Ontario, Canada; 6 Li Ka Shing Knowledge Institute, St. Michael’s Hospital, Canada; 7 Division of Pharmacoepidemiology and Pharmacovigilance, Spanish Medicines and Healthcare Products Agency (AEMPS), Spain; 8 The First Clinical College of Lanzhou University, China; 9 Department of Pharmacy, Vancouver General Hospital, Canada; 10 Federal University of Pelotas, Brazil Background: Moher et al. [1] examined a November 2004 MEDLINE-indexed sample of 300 systematic reviews (SRs) and found that reporting quality was variable. The SR landscape has changed considerably in the subsequent decade, with the publication of a reporting guideline for SRs of randomised trials of health care interventions (PRISMA), methodological developments such as the Cochrane risk Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. RO 4.7 Challenges and considerations in assessing the reporting quality of systematic reviews in overviews using PRISMA Jia P1 , Chen J2 , Zhang L1 , Zhao P1 , Zhang M1 1 Chinese Cochrane Center, China; 2 West China Hospital, China Background: Overviews bring systematic reviews (SRs) together and serve as a user-friendly ’digest’ that can be used by clinicians and policy makers in making decisions. Poor reporting of SRs diminishes value of the overview to clinicians and policy makers. The reporting quality of SRs can be evaluated using the PRISMA (Preferred Reporting Items for Systematic reviews and Meta-Analysis) system. Objectives: To evaluate the reporting quality of SRs involved in overviews; to examine the difference exiting PRISMA scores make across Cochrane and non-Cochrane SRs. Methods: We randomly selected eight overviews and searched all the SRs included in the overviews. Each PRISMA item is given a score of 1 if the criterion is met, a score of 0.5 if the criterion is partially met, or a score of 0 if the criterion is not met, or not applicable. Two reviewers independently evaluated the quality of included SRs and discrepancies were solved by Cochrane Database Syst Rev Suppl 1–327 (2015) 23 consensus. Results: Ninety-eight SRs were included; 28 of these were published in 2009. The 98 SRs were from four major areas: cancer, hypertension, skin and stroke. Among the 98 SRs, 36 SRs were Cochrane SRs. The total PRISMA scores were (22.5 ± 10.2). Average Cochrane scores (25.5 ± 11.7) were much higher than the non-Cochrane reviews (20.5 ± 13.5). The reporting quality varied in both Cochrane and non-Cochrane reviews: several items were strong, including reporting of data collection process (100%), selection of articles (100%) and synthesis of results (100%). In some items, the reporting quality of Cochrane SRs was better than the non-Cochrane SRs, such as: protocol and registration and reporting of search strategies. Conclusions: The reporting quality of Cochrane SRs was higher than that of non-Cochrane SRs. The missing information in the Cochrane and non-Cochrane SRs may influence the use of overviews, so the reporting quality in the Cochrane and non-Cochrane SRs needs to be improved. Rapid oral session 5 Online databases for evidence dissemination RO 5.1 Alessandro Liberati online Library: a tool to increase public health professionals’ knowledge and practices Vecchi S1 , De Fiore L2 , Amato L3 , Davoli M3 1 Department of Epidemiology, Lazio Regional Health Service, Italy; 2 Associazione Alessandro Liberati/Network Italiano Cochrane, Italy; 3 Department of Epidemiology, Lazio Regional Health Service; Cochrane Drugs and Alcohol Group, Italy Background: Public health professionals are increasingly expected to use research evidence in decision making. Over the last 10 years, there has been an increasing investment to identify barriers to incorporate evidence in practice such as the growing volume of literature, which forces health professionals to adapt their knowledge and skills to the rapid development of both medical and scientific research and technical innovation. Some interventions aimed at addressing this issue often do not take the real information needs of the health personnel, nor the quality and reliability of information sources accessed, into account. Objectives: To evaluate whether the integration of an open access to evidence source with a tailored education program improves adherence to evidence-based clinical practices among health personnel in the Lazio Region. Methods: A two-phase approach was used: 1. development of a regional online Medical Library (Biblioteca Alessandro Liberati/BAL) in order to provide an organized and well-structured bibliographic resource enabling health professionals quick access to evidence-based information; 2. pilot testing of an educational programme aimed at empowering health personnel to find the information of interest, gaining knowledge of databases and search engines and training them on the critical appraisal of published Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. studies. Participants have been engaged in interactive workshops, formal lessons and common discussion on the appropriate use of evidence through the BAL. An 11-item questionnaire was developed and administered to explore EBM attitudes and skills before the intervention. We planned to administer the questionnaire 12 months after the intervention to evaluate the improvement in skills for using evidence sources, ability in searching and retrieval of high quality information. Results: BAL has been available online since May 2014. To date, four local health units in the Lazio Region have been involved in the pilot test. Data collection and analysis will be complete for presentation at the 2015 Colloquium. RO 5.2 Patient portals for enhancing patient empowerment Lannig S1 , Ammenwerth E1 , Schnell-Inderst P1 , Hörbst A1 1 UMIT Private University for Health Sciences, Medical Informatics and Technology, Austria Background: Patient portals are applications that allow patients to access patient-related health information that is documented by a healthcare institution electronically. As part of a patient portal, institutions may offer patients (typically web-based) access to selected clinical data that is governed by the institutions as part of the electronic health record (EHR). The patient can access this clinical data, read and print it, or integrate it into any (electronic or paper-based) patient-owned record. Thus, patients are supported to manage their own health care by using these patient portals. Patient empowerment, patient satisfaction and other health-related outcomes may be affected by using patient portals. Already in 1986, the Ottawa Charter for Health Promotions had made empowerment a key issue in the theory of health promotion, but only a few controlled studies on the impact of patient portals are available at the moment. There seem to be large uncertainties regarding the impact of patient portals on patient empowerment. Objectives: To conduct a systematic review of the effect of patient portals on patient empowerment as a patient-related outcome. Methods: We will conduct a systematic literature search on patient portals. We will search in health-related databases like MEDLINE, CINAHL, and Academic Search Elite. In addition we will handsearch for literature in journals of medical and nursing informatics. Subsequently two independent scientists will screen all located abstracts. Initially we will include randomized controlled trials. After that we will assess the quality of these studies based on the Cochrane ’Risk of bias’ tool. We will summarize the findings related to the impact of patient portals especially for consumers. Expected results: We expect to find a small number of patient-related outcomes addressing the impact of patient portals. Probably, there will be insufficient data to make a general recommendation for implementing patient portals, but we may find an impact for special sub-types such as portals for patients with diabetes. Cochrane Database Syst Rev Suppl 1–327 (2015) 24 RO 5.3 Cochrane Clinical Answers: filtering the information overload for better clinical decisions Tort S1 , Pettersen K2 , Burch J1 1 Cochrane Editorial Unit, United Kingdom; 2 Wiley, United Kingdom Background: Healthcare professionals need point-of-care access to reliable and high-quality synthetized information from up-to-date research. Cochrane Reviews provide a valuable source of information, but can take a long time to read and have a large volume of data, much of which may not be relevant to the busy healthcare professional. Cochrane Clinical Answers (CCAs) provide an accessible, clinically-focused summary of Cochrane Reviews and therefore can support decision making at the bed-side. Objectives: To describe the current and proposed coverage of CCAs, and how their presentation aids information dissemination. Methods: The CCA website was developed to mimic the way clinicians approach information gathering. To ensure appropriate coverage, the number of published CCAs by discipline was reviewed and targets established for 2015. Results: Each CCA addresses a question and provides a concise answer. Full outcome data supporting the answer are a click away. The population, intervention and comparator (PICO) information, a narrative result, the quality of evidence or risk of bias summary, a link to the forest plot and absolute values are provided to allow quick understanding and application of results. By February 2015, we had published 645 CCAs across all disciplines. Coverage ranged from 7% (in the areas of Neonatal, and Pregnancy and Childbirth) to 89% (Metabolic and Endocrine Disorders), with Oncology and Infectious Diseases being made priorities in 2015. Six-hundred CCAs were published by January 2015, and our target for January 2016 is 1000. New developments involve creating CCAs from overviews, network meta-analyses, and narrative reviews, as clinicians would benefit most from reliable and accurate clinician-friendly summaries of these complex Cochrane Reviews. Conclusions: CCAs are a great tool to filter the vast amount of evidence from Cochrane Reviews and make it easier for healthcare professionals to apply high-quality evidence when managing patients. RO 5.4 Introducing KSR Evidence: developing a database of critically appraised systematic reviews in the field of pain relief Misso K1 , Deshpande S1 , Stirk L1 , Westwood M1 , Kleijnen J1 , McLellan A1 , Kleijnen J1 1 Kleijnen Systematic Reviews Ltd, United Kingdom Background: Publication of systematic reviews (SRs) has increased dramatically in recent years, from 232 in 1990 to 11,146 in 2014 (PubMed estimate). Assessing quality, risk of bias and applicability of review findings remains a challenge Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. for clinicians and practitioners. KSR Evidence will include critical appraisals of systematic reviews from 2010 onwards. KSR Pain Evidence, as part of KSR Evidence, will enable quick and easy access to current critical appraisals of published SRs in pain management, and assist in interpretation and application of findings into clinical practice. Objectives: To develop a database that comprehensively covers the best systematic review evidence in pain management, together with related randomised controlled trials (RCTs) from PubMed. The database includes critical appraisals for each SR, with a summary of SR results and a concise clinically-relevant bottom line. Methods: Sensitive searches were conducted on a wide range of resources, including MEDLINE, Embase and the Cochrane Database of Systematic Reviews (CDSR), to retrieve systematic reviews on pain management from 2010 onwards (see Figure 1). For each included SR, a detailed critical appraisal and risk of bias was undertaken by an expert reviewer, culminating in a clinical bottom line. Appraisals were conducted using an adaptation of the ROBIS (Risk of Bias In Systematic reviews) tool, and are checked for quality by a second independent reviewer. Results: To date, KSR Pain Evidence comprises over 2456 critical appraisals, and 2183 related RCTs from PubMed. We will present KSR Pain Evidence, describe the rationale behind its development and the appraisal process. Search functionality and example appraisals will be demonstrated. Conclusions: Further work is ongoing to expand database coverage into other disease areas, and it is anticipated that KSR Evidence will extend to all published systematic reviews in the near future (www.ksrevidence.com). Cochrane Database Syst Rev Suppl 1–327 (2015) 25 RO 5.5 CBRG QuickDecks: a new tool for sharing the best evidence in back and neck pain care Irvin E1 , Furlan A1 , Munhall C1 , Danak S1 , Kim J2 , Marin T3 1 Cochrane Back Review Group, Canada; 2 Cancer Care Ontario, Canada; 3 York University, Canada Background: It is important to create different summary products in addition to scientific publications to support claims about interventions based on evidence from Cochrane Reviews. The Cochrane Back Review Group (CBRG) wanted to create a product that would decrease the burden and yet encourage the use of evidence from Cochrane Reviews for busy researchers and clinicians. Objectives: To create a product to share findings from reviews published within the CBRG that provides a quick snapshot of the evidence in an accessible format. Methods: QuickDecks are created for each intervention review and classified into eight different intervention themes. They consist of three summary slides with a standard format presenting objectives, methods, results and conclusions of the review. QuickDecks, prepared by the CBRG, are sent to the review authors for approval. PDF versions of the slides are available for anyone to download and PowerPoint versions are available upon request. Results: QuickDecks have been promoted through the CBRG Twitter, Facebook, newsletters and email communications. Since the QuickDecks have been available to download in August 2014, there has been a lot of positive feedback and a large number of downloads. Conclusions: QuickDecks are a new tool used to share evidence and promote uptake of Cochrane review findings. The results of this submission may be viewed at: https://ccnc.cochrane.org/node/605/submission/5975 RO 5.6 IN SUM: mapping the evidence on effects of child mental health and welfare interventions Hammerstrøm K1 , Ludvigsen K1 , Axelsdottir B1 , Biedilæ S1 , Kornør H1 1 Centre for Child and Adolescent Mental Health, Eastern and Southern Norway (RBUP), Norway Background: Systematic reviews (SRs) on the effect of interventions are important sources of research for practitioners and other stakeholders in the fields of child welfare and child mental health. However, finding, appraising and using research literature may appear both difficult and time-consuming to practitioners. Objectives: To support practitioners’ use of research by identifying and quality appraising all SRs on the effect of interventions in child welfare and child mental health. Methods: To identify SRs investigating the effect of any intervention concerning child (< 18 years) welfare or child mental health we carried out pilot searches in bibliographic databases (the Cochrane and Campbell Libraries, PsycINFO, MEDLINE, EMBASE, DARE etc.) focusing on specific topics of interest (e.g. child Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. depression and anxiety;). However, the overlaps between the subject specific searches were considerable. For this reason, we decided to carry out one broader, exhaustive search per database, in order to identify all potentially relevant SRs simultaneously. The searches will be re-run semi-annually. Inclusion/exclusion: Two researchers will independently screen identified titles/abstracts and quality assess included SRs, using AMSTAR (A MeaSurement Tool to Assess systematic Reviews). Following inclusion, references will be coded and published in the IN SUM database, which is under development - but will be accessible, and free, for all. For selected SRs, we will produce Norwegian/English plain language summaries. Where we identify synopses of syntheses (e.g. DARE or Health Evidence summaries) for included SRs, we will provide a link to the synopsis in the relevant SR reference. Results: The exhaustive search will be conducted in March 2015. We estimate that we will complete the initial selection, appraisal and publication process by 1 July 2015. Conclusions: SRs are important in informing decisions in child mental health and welfare. We believe that through the provision of easy access to quality-appraised relevant SRs and plain language summaries, the IN SUM database has the potential of becoming a valuable source for practitioners in our field. RO 5.7 Building a database of respiratory guidelines: why, how, and what for? Welsh E1 , Stovold E1 1 Cochrane Airways Group, United Kingdom Background: Cochrane Review Groups are encouraged to demonstrate impact by tracking guidelines that incorporate their reviews. To help with this we began to compile a collection of guidelines that cite Cochrane Airways Group reviews. This grew into the idea of establishing a general collection of respiratory guidelines, including not only the guidelines that cite our reviews, but other relevant guidelines that may be helpful to our review authors and others with an interest in respiratory medicine. Objectives: To create an open database of respiratory guidelines. Methods: We began with a known set of guidelines acquired through general knowledge of the topic area. We set up a monthly search alert in PubMed. The search was set up in January 2015, and run back over 2014. Older guidelines were identified by checking major respiratory society websites. We use Zotero Groups to manage the database. It is a public group, with closed membership. Anyone can view the records, but membership is required to edit the database. The database was promoted through social media, and is linked on the Airways Group website. Results: We structured the database around disease categories by creating individual collections. The database of guidelines includes asthma, wheeze, COPD (chronic obstructive pulmonary disease), bronchiectasis, obstructive sleep apnoea, chronic cough, interstitial lung diseases, and pulmonary hypertension. We also have collections for emergency care and physiotherapy. Cochrane Database Syst Rev Suppl 1–327 (2015) 26 Conclusions: We established a reference collection of respiratory guidelines, free for all to view. It has been compiled through pragmatic methods using a simple search strategy and is not an exhaustive collection. We would like to develop this database further, get more people involved in contributing guidelines, and seek feedback. RO 5.8 Best evidence for optimal aging: filter the information overload using the McMaster Optimal Aging Portal Dobbins M1 , Watson S2 1 McMaster University, Canada; University, Canada 2 Health Evidence, McMaster Background: There is so much health information freely available on the internet relevant to ’optimal aging’ (staying healthy, active and engaged as we grow older), that it is difficult to know which messages are trust-worthy. The McMaster Optimal Aging Portal does the work for you: evaluating, summarizing and synthesizing aging research and resources to provide you with a one-stop shop for evidence-based information on healthy aging. Objectives: The McMaster Optimal Aging Portal is a website dedicated to sharing evidence-based information about optimal aging and common health conditions with aging adults, their caregivers and health professionals. Funded by the Labarge Optimal Aging Initiative, the website was launched in October 2014. Methods: Content is drawn from three best-in-class resources: MacPlus (a database of best evidence to support clinical decisions); Health Evidence (an online registry of quality-appraised systematic reviews evaluating public health interventions); and Health Systems Evidence (an online registry of evidence on how to strengthen or reform health systems). Star-ratings allow users to filter through masses of health information and research and easily identify the highest quality resources they need to make evidence-based decisions around optimal aging. Health professionals can access abstracts and summaries of the best scientific evidence on clinical, public health and health policy questions, including links to full-text articles. Blog posts, web resource ratings and evidence summaries provide quick and easy-to-read ‘bottom line’ messages appropriate for all audiences, particularly citizens. Tailored email alerts and @Mac AgingNews updates deliver the latest research evidence on headline topics right into your hands. Results and Conclusions: The McMaster Optimal Aging Portal gives citizens, health professionals, policy makers and researchers direct access to trusted, evidence-based information about optimal aging to help us remain healthy, active and engaged as long as possible. RO 5.9 Coming full circle: returning to a bespoke search approach for clinical evidence Beaven O1 1 BMJ Group, United Kingdom Background: Clinical Evidence (CE) is a long running product from BMJ Group, which has been providing systematic overviews of the best available research evidence on selected interventions for numerous medical conditions for over 10 years. During this time CE has developed from print only, to CD-ROM and on to its current online format and has had a number of changes in process over this time period - one such component being the approach taken to searching the literature. Objectives: To reflect on how the literature searching function has developed over time to support the production of CE – specifically its progress from bespoke searches, to a broad multi-product, condition-level approach, and currently, to its ongoing return to specific intervention-based search strategies. Methods: A personal perspective, considering the advantages and disadvantages of the searching approaches that have been applied, the resources that have been utilised, the constraints and limitations that have been negotiated and how we continue to strive to get the balance right, in order to provide the quality of search required to support an evidence-based, regularly updated, commercial product. Results and Conclusions: There has been a great deal of change over the lifetime of CE and different methods have been applied to try to maintain quality whilst expanding, enhancing and developing a useful and relevant product for end users. Whilst the results and conclusions are not yet formulated, I would hope that the insight into the approaches that we have undertaken will provide some interesting observations, different ideas, and perhaps stimulate some discussion amongst other information professionals about how to deliver pragmatic searches for continually evolving products. Rapid oral session 6 Individual patient data RO 6.1 A framework for deciding if individual participant data are likely to be worthwhile Tudur Smith C1 , Clarke M1 , Marson T2 , Riley R1 , Stewart L1 , Tierney J1 , Vail A3 , Williamson P1 1 IPD Meta-analysis Methods Group, United Kingdom; 2 Cochrane Epilepsy Group, United Kingdom; 3 Cochrane Menstrual Disorders & Subfertility Review Group, United Kingdom Background: Systematic reviews and meta-analysis that are based on Individual Participant Data (IPD) are often referred to as a ’gold standard’. However, they are usually Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Cochrane Database Syst Rev Suppl 1–327 (2015) 27 more resource-intensive than a traditional approach based on aggregate data (AD) extracted from publications or made available from authors. The advantages and disadvantages of the IPD approach have been discussed extensively in the literature. In many cases the decision about whether to collect IPD can be straightforward. For example, if suitable AD are not available in the trial publications, some data will be needed from the original researchers if the trial is to be included in any meta-analyses. However, in some cases it can be difficult for a researcher, or research funder, to decide whether the extra investment of time and money will be worthwhile. Further guidance would be helpful to assist with this decision making. Objectives: To provide a framework for researchers, and research funders, to help decide whether collecting IPD would be worthwhile. Methods: Existing empirical evidence comparing IPD and AD, literature describing the advantages and disadvantages of IPD, and the expertise of the Cochrane IPD Meta-analysis Methods Group will be used to develop an IPD decision framework. Results: The framework will be described and made available at the Colloquium. Conclusions: The framework will help researchers and funders determine when IPD is likely to be most beneficial. This could avoid resources being wasted on unnecessary IPD reviews, help to achieve the maximum potential of IPD, and help to prioritise research projects that would benefit most from IPD. RO 6.2 Imputation of systematically missing predictors in an individual participant data meta-analysis: a generalized approach using MICE Debray T1 , Jolani S2 , Koffijberg H1 , van Buuren S2 , Moons K1 1 Julius Center for Health Sciences and Primary Care, The Netherlands; 2 Utrecht University, The Netherlands Background: Individual participant data meta-analyses (IPD-MA) are increasingly used for developing and validating multivariable (diagnostic or prognostic) risk prediction models. Unfortunately, some predictors or even outcomes may not have been measured in each study and are thus systematically missing in some individual studies of the IPD-MA. As a consequence, it is no longer possible to evaluate between-study heterogeneity and to estimate study-specific predictor effects, or to include all individual studies, which severely hampers the development and validation of prediction models. Objectives: To describe a novel approach for imputing systematically missing data, which adopts a generalized linear mixed model to allow for between-study heterogeneity. Methods: We illustrate our approach using a case study with IPD-MA of 13 studies to develop and validate a diagnostic prediction model for the presence of deep venous thrombosis. We compare the results after applying four methods for dealing with systematically missing predictors in one or more individual studies: complete case analysis (CCA) where studies with systematically missing predictors are removed, traditional multiple imputation Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. ignoring heterogeneity across studies (TMI), stratified multiple imputation accounting for heterogeneity in predictor prevalence (SMI), and multilevel multiple imputation (MLMI) fully accounting for between-study heterogeneity. Results: Results from CCA were suboptimal and became completely unreliable when the predictors were no longer missing completely at random. TMI and SMI tended to mask the actual degree of between-study heterogeneity and often lead to overoptimistic standard errors of predictor effects. MLMI was the optimal approach in terms of coverage and bias, and the only approach that was able to ensure compatibility of imputation and analysis models. Conclusions: MLMI may substantially improve the estimation of between-study heterogeneity parameters and allow for imputation of systematically missing predictors in IPD-MA aimed at the development and validation of prediction models. RO 6.3 The advantages of individual participant data (IPD) in smaller-scale systematic reviews Scott N1 , Spiteri Cornish K2 , Lois N3 1 Medical Statistics Team, University of Aberdeen, United Kingdom; 2 Ophthalmology Department, NHS Grampian, Aberdeen, United Kingdom; 3 Centre for Experimental Medicine (CEM), Queen’s University, Belfast, United Kingdom Background: Individual participant data (IPD) have been used in many systematic reviews of randomised controlled trials. Commonly-cited benefits include the ability to perform meta-analyses of time-to-event outcomes, the ability to undertake subgroup analyses and to ensure data quality. Many published IPD reviews have involved large collaborative groups and large numbers of studies. We report the unique benefits of IPD in a small-scale review in ophthalmology. Methods: Systematic review of studies comparing internal limiting membrane (ILM) peeling during macular hole surgery versus no peeling. Results: IPD were received from the three identified fully-published studies. The authors of a fourth study were not willing to provide data until full publication but limited data were available from an abstract. The review was able to answer key questions about the effectiveness of ILM peeling. Benefits of IPD in this review included the ability to standardise the unit of randomisation (eye or patient), to include only randomised participants in one study with an unusual format, to explore timing of assessments, to include information from both IPD and published data, to conduct subgroup analyses by stage of macular hole and to conduct adjusted analyses controlling for study, size and duration of the macular hole. Conclusions: This was a successful collaboration and a review of published data alone would have been very difficult. The benefits, however, were very different to those in a previous large-scale IPD review conducted by the primary author. In particular, the ability to standardise the unit of randomisation was crucial to the success of this review. This project demonstrates that there may be a variety of reasons to seek IPD, even in smaller systematic reviews. Cochrane Database Syst Rev Suppl 1–327 (2015) 28 RO 6.4 Development of a unified database based on CDISC for individual patient data (IPD) analysis: practices in reproductive health survey Zhang Y1 , Fei Y1 , Li X1 , Liu J1 1 Center for Evidence Based Chinese Medicine, Beijing University of Chinese Medicine, China Background: IPD meta-analysis is suggested as the least biased method, which might provide more detailed and reliable conclusions than conventional meta-analysis. Since the original data from individual studies can be re-categorized, IPD meta-analysis can provide more consistent results. Before initiating an IPD analysis, establishing comprehensive data standards is essential to the development of a database that enables the pooling of data from different sources. The Clinical Data Interchange Standards Consortium (CDISC) standards are preferred by regulators, industry, and other research organizations as a means of facilitating regulatory review, aggregation, and querying of data, sharing data between entities, and streamlining the acquisition and analysis of data. Objectives: We aimed to develop a unified database to standardize and collect data from various sources. Methods: Study Data Tabulation Model (SDTM) standard is a foundational dataset for data management procedures. SDTM is suited for collecting data of various types and storing it in a relatively small number of observation classes. We examined the data structures of three epidemiology investigations, and then developed new data domains for storing reproductive-specific data. With the standards in place, patient-level data from the three studies were remapped and used to construct the database. Logical mapping and programmatic transformation were applied in the whole process. Validating and reconciling were made following related standard operating procedures. Results: To date, the unified database contains six domains (DM, SC, CM, MH, VS, QS) and has received data on a total of 23600 subjects from three remapped studies of reproductive health submitted by two researchers in the National Research Institute for Family Planning. Conclusions: Developing data standards is essential for IPD meta-analysis. A unified database would facilitate the pooling of data from legacy studies. In addition, it is also a data reservoir for prospectively collecting data in new trials without the need for remapping. RO 6.5 Methods and characteristics of published network meta-analyses using individual patient data Veroniki AA1 , Soobiah C1 , Tricco A1 , Elliott M1 , Straus SE1 1 Li Ka Shing Knowledge Institute, St. Michael’s Hospital, Toronto, Canada Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Background: Individual patient data (IPD) may increase precision in network meta-analysis (NMA) compared to using aggregated data (AD), as it utilizes both within-study and across-study data. The use of IPD in NMA also prevents misleading inferences due to bias, such as aggregation bias, and allows for the identification of interactions due to patient-level effect modifiers, that cannot be detected using AD. While applications of IPD-NMA methods have been published, a scoping review presenting all approaches to applying an IPD-NMA has not been conducted. Objectives: To synthesize existing IPD-NMA methods, summarize their properties, and describe the characteristics of published IPD-NMAs. Methods: We searched relevant electronic databases from inception until September 2014 and for grey literature through Google. Published and unpublished studies were eligible, including studies suggesting and/or applying IPD-NMA methods or studies addressing the methodological quality of IPD-NMA. We included IPD-NMAs of all quantitative study designs. Several approaches have been presented in the literature for modeling IPD alone or in combination with AD in NMA, which vary in popularity and complexity. We will provide guidance for selecting an appropriate IPD-NMA model, identify gaps where methodology is lacking, and highlight the potential for novel statistical advances which are necessary to evaluate the key assumptions in NMA. We will present the characteristics of previously conducted IPD-NMAs, and the methods that have been applied. Quantitative data (e.g. number of patients) will be summarized by medians and interquartile ranges, while categorical data (e.g. effect size) will be summarized by numbers and percentages. Results: The results will be ready by September 2015 and will be presented at the Cochrane Colloquium. Discussion: Our study will provide a comprehensive overview of the methods for completing an IPD-NMA, as well as insight into the characteristics of IPD-NMAs that have been conducted. RO 6.6 Complex interventions and composite outcomes in Individual Participant Data (IPD) meta-analysis: double challenge in i-WIP IPD Rogozinska E1 , Marlin N2 , Thangaratinam S1 1 Women’s Health Research Unit, Queen Mary University of London, United Kingdom; 2 Pragmatic Clinical Trials Unit, Queen Mary University of London, United Kingdom Background: i-WIP (International Weight management In Pregnancy) is an IPD meta-analysis of randomised controlled trials looking at the differential effect of diet and physical activity based interventions on maternal and fetal outcomes in pregnancy. The interventions used in the studies eligible for the IPD involve a range of weight management methods aiming to improve diet, level of physical activity and lifestyle. The interventions were delivered with various frequencies, for different lengths of time and formats (group sessions and individual consultations). Additionally the effect of interventions was evaluated using a wide range of maternal, Cochrane Database Syst Rev Suppl 1–327 (2015) 29 fetal and neonatal outcomes. Objective: To discuss the solutions adopted in i-WIP project to deal with a double challenge of complexity of evaluated interventions with a multitude of clinical outcomes. Complex interventions: Building on the work of Hill et al. (2013), we developed an 18-item questionnaire (Table 1) breaking down the intervention into standard components. We asked the trials’ Principal Investigators (PIs) to fill out the questionnaire and specify which of the items were present in their interventions. Outcomes: We conducted a two-stage Delphi survey to prioritize the outcomes for their importance to clinical care. Secondly, using a validated method we developed maternal Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. composite, and fetal and neonatal composite outcomes (Table 2). Discussion: The IPD meta-analysis gives the researchers greater power than the aggregated one, however it comes with certain challenges. If our abstract is accepted, we will provide a detailed description of the solutions applied in the i-WIP IPD; their strengths, limitations and recommendation for future research. Reference: Hill B., Skouteris H. and Fuller-Tyszkiewicz. Interventions designed to limit gestational weight gain: a systematic review of theory and meta-analysis of intervention components. Obesity Reviews 2013; 435–450. Cochrane Database Syst Rev Suppl 1–327 (2015) 30 RO 6.7 Selective cutoff reporting in diagnostic test accuracy studies of depression screening tools: comparing traditional meta-analysis to individual patient data meta-analysis Levis B1 , Benedetti A1 , Thombs B1 1 McGill University, Canada Background: Selective outcome reporting in clinical trials is well understood, but has not been assessed systematically in studies of diagnostic test accuracy, where authors often report results for a small range of cutoffs around data-driven ’optimal’ cutoffs maximizing sensitivity and specificity. Objectives: To compare traditional meta-analysis of published results to individual patient data (IPD) meta-analysis of results from all cutoffs, to: (1) assess the degree to which selective cutoff reporting exaggerates accuracy estimates; and (2) identify patterns of selective cutoff reporting. Methods: Bivariate random-effects models were used to compare results of traditional and IPD meta-analysis, using studies included in a published meta-analysis of the Patient Health Questionnaire-9 (PHQ-9) depression-screening tool (Manea, CMAJ, 2012). Results: Thirteen of 16 primary datasets were obtained. For the ’standard’ cutoff of 10, most studies (11 of 13) published accuracy results. For all other cutoffs, only three to six of the 13 studies published accuracy results. For all cutoffs, specificity estimates in traditional and IPD meta-analyses were within 2%. Sensitivity estimates were similar for cutoff 10, but differed by 5% to 15% for all other cutoffs. In samples where the PHQ-9 was poorly sensitive, authors reported results for cutoffs around the low optimal cutoff. In samples where the PHQ-9 was highly sensitive, authors reported results for cutoffs around the high optimal cutoff. Consequently, in the traditional meta-analysis (but not in the IPD meta-analysis), sensitivity increased as cutoff severity increased for part of the range of possible cutoffs. Comparing cutoff 10 across all studies, sensitivity was heterogeneous (tau-squared = 1.95). Comparing optimal cutoffs, however, sensitivity was more homogeneous (tau-squared = 0.68), but cutoff values ranged from 5 to 15. Conclusion: Selectively reporting well-performing cutoffs in small samples leads to biased estimation of accuracy in traditional meta-analyses. To reduce bias in meta-analyses, primary studies should report accuracy results for all cutoffs. Reference: Manea L, Gilbody S, McMillan D. Optimal cut-off score for diagnosing depression with the Patient Health Questionnaire (PHQ-9): a meta-analysis. CMAJ 2012; 184: E191–6. Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Lunch rapid oral session 1 Minimally important difference LRO 1.1 Minimally important difference estimates for patient reported outcomes in pediatrics: a systematic survey Ebrahim S1 , Nesrallah G2 , Vercammen K3 , Sivanand A3 , Carrasco-Labra A3 , Furukawa T4 , Guyatt G3 , Johnston B5 1 McMaster University; Hospital for Sick Children, Canada; 2 Li Ka Shing Knowledge Institute, St. Michael’s Hospital, Canada; 3 McMaster University, Canada; 4 Kyoto University, Canada; 5 Hospital for Sick Children, Canada Background: The minimally important difference (MID) provides a measure of the smallest change in a patient reported outcome (PRO) instrument that patients perceive as important. An anchor-based approach is widely considered the most appropriate method for MID determination. No study or database currently exists that documents and appraises all anchor-based MIDs associated with PRO instruments in children. Objectives: To complete a systematic survey of the literature to collect, characterize and appraise published anchor-based MIDs associated with PRO instruments used in evaluating the effects of interventions on chronic medical and psychiatric conditions in children. Methods: We searched MEDLINE, EMBASE, and PsycINFO (1989 to 11 February 2015) to identify studies reporting empirical ascertainment of anchor-based MIDs in pediatric PROs. Teams of two reviewers screened titles and abstracts, reviewed full texts of citations identified as potentially eligible, extracted relevant data, and assessed the credibility of all studies. Results: Of 3910 citations, 706 were reviewed in full text; 32 proved eligible. Anchor-based MID determinations were assessed in 23 (72%) disease-specific PROs and 10 (28%) generic PROs (one study reported both). Fourteen PROs and anchors were self-reported, 10 PROs and 13 anchors were proxy-reported (parent or guardian), eight PROs and three anchors were both self- and proxy-reported, and reporting of two anchors was unclear. Anchor-based approaches consisted of a standard approach (range of scores on an independent anchor that was classified as ’minimally important change’) in 27 studies, three used Receiver Operating Characteristic curves, and two used an approach based on hypothetical scenarios. Detailed results and the credibility assessments of the anchor-based approaches will be presented at the Cochrane Colloquium. Conclusions: We have compiled the first comprehensive synthesis of anchor-based MIDs for PROs in children. This compendium will help clinical trialists in child health improve sample size calculations, and will enhance the interpretability of treatment effects in systematic reviews and practice guidelines. Cochrane Database Syst Rev Suppl 1–327 (2015) 31 LRO 1.2 Minimally important difference estimates and assessment of its credibility for patient reported outcomes in adults: a systematic survey Carrasco- Labra A1 , Johnston BC2 , Ebrahim S2 , Furukawa TA3 , Patrick DL4 , Hemmelgarn BR5 , Motaghi Pisheh S1 , Dennis B1 , Jacobs M6 , Vernooij RW7 , Fei Y8 , Younho Hong B1 , Tajika A3 , Takeshima N3 , Iwakami N3 , Hayasaka Y3 , Kaminski-Hartenthaler A9 , Nussbaumer B9 , Colunga L10 , Gartlehner G9 , Schunemann HJ1 , Guyatt G1 , Nesrallah G11 1 Department of Clinical Epidemiology & Biostatistics, McMaster University, Canada; 2 Department of Anaesthesia and Pain Medicine, The Hospital for Sick Children; Department of Clinical Epidemiology & Biostatistics, McMaster University, Canada; 3 Department of Health Promotion and Human Behavior, School of Public Health, Kyoto University Graduate School of Medicine, Japan; 4 Seattle Quality of Life Group, Department of Health Services, University of Washington, USA; 5 Faculty of Medicine, University of Calgary, Canada; 6 Department of Medical Psychology J3-220-1, Academic Medical Center, The Netherlands; 7 Iberoamerican Cochrane Centre, Institute of Biomedical Research (IIB Sant Pau), Spain; 8 Centre for Evidence-Based Chinese Medicine, Beijing University of Chinese Medicine, Department of Clinical Epidemiology & Biostatistics, McMaster University, China; 9 Department for Evidence-Based Medicine and Clinical Epidemiology, Danube University, Krems, Austria; 10 Hospital Civil de Guadalajara, Fray Antonio Alcalde, Mexico; 11 Nephrology Program, Humber River Regional Hospital; Division of Nephrology, University of Western Ontario, Canada Background: Patient-reported outcomes (PROs) have the function of reflecting patients’ perspectives on treatment benefits and harms. The ability to interpret and, consequently, include PROs in the decision-making process along with other patient-important outcomes relies on the understanding of what improvement or reduction in the PRO represents a change that is considered important to patients. The most common reference point for interpretation of PROs is the minimal important difference (MID), which provides a measure of the smallest change in a PRO that patients would perceive as an important benefit or harm. Objective: To document published anchor-based MIDs associated with PRO instruments used in evaluating the effects of interventions on chronic medical and psychiatric conditions in both adult and pediatric populations. Methods: We are currently launching a study that summarizes and appraises available methods to estimate anchor-based MIDs. We will also document all anchor-based MIDs reported for adults and pediatric populations and will develop a rating instrument to assess their credibility. We will create an inventory of available anchor-based MIDs, including the context in which they were assessed (condition/disease), and the confidence users can place in a particular MID. We will search MEDLINE, EMBASE, PsycINFO, and CINAHL (1989 to present) to identify studies addressing methods to estimate anchor-based MIDs of PRO instruments or reporting empirical ascertainment of anchor-based MIDs. Teams of two reviewers will Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. independently screen citations identified, and extract relevant data. Discussion: No inventory of MIDs is currently available, which means that clinicians and patients have to navigate the vast literature in order to retrieve a specific MID. Even if they find an apparently applicable MID, there is no guidance regarding the extent to which users can trust the reported MID. We will provide a summary of the available MIDs in the literature and describe their overall credibility. LRO 1.3 Though not consistently superior, the absolute approach to framing the minimally important difference has advantages over the relative approach Zhang Y1 , Zhang S1 , Thabane L1 , Furukawa TA2 , Johnston BC3 , Guyatt GH4 1 Department of Clinical Epidemiology and Biostatistics, McMaster University, Canada; 2 Department of Health Promotion and Human Behavior, Kyoto University School of Public Health, Kyoto, Japan; 3 Department of Anaesthesia and Pain Medicine, The Hospital for Sick Children, University of Toronto; Institute of Health Policy, Canada; 4 Department of Clinical Epidemiology and Biostatistics, and Medicine, McMaster University, Canada Background: Whether the minimal important difference (MID) is best framed using an absolute change or a relative change in score when interpreting patient-reported outcomes remains uncertain. Objectives: Using studies that established MID using anchor-based methods, we set out to address the relative merits of absolute and relative changes in establishing an instrument’s MID. Methods: In seven data sets, we calculated correlations between global change ratings and absolute and relative score changes and conducted meta-analyses. We considered that the measure with the higher correlation represented the more valid approach. Results: The meta-analyses showed no significant difference between pooled correlations of absolute and relative difference on a health-related quality of life instrument with global transition scores of symptoms, emotional function, physical function, and cognitive function. In four of five domains there was at least one study in which the absolute was significantly superior to the relative; in one of these four, one study showed statistically significant superior performance of the relative. In an analysis restricted to patients with low baseline scores for the domain of cognitive function, the relative approach showed higher correlation with global rating than did the absolute approach. Conclusions: Although we found no consistent superiority of either approach to establishing the MID, when differences existed they usually favoured the absolute, which also has advantages of simplicity and ease of pooling across studies. Researchers may consider the absolute as a default, but also compare both methods on an instrument by instrument basis. Cochrane Database Syst Rev Suppl 1–327 (2015) 32 LRO 1.4 A new tool to measure credibility of studies determining minimally important difference estimates for patient reported outcomes Carrasco- Labra A1 , Johnston BC2 , Ebrahim S2 , Furukawa TA3 , Patrick DL4 , Hemmelgarn BR5 , Motaghi Pisheh S1 , Dennis B1 , Jacobs M6 , Vernooij RW7 , Fei Y8 , Younho Hong B1 , Tajika A9 , Takeshima N9 , Iwakami N9 , Hayasaka Y9 , Kaminski-Hartenthaler A10 , Nussbaumer B10 , Colunga L11 , Gartlehner G10 , Schunemann HJ1 , Nesrallah G12 , Guyatt GH1 1 Department of Clinical Epidemiology & Biostatistics, McMaster University, Canada; 2 Department of Anaesthesia and Pain Medicine, The Hospital for Sick Children; Department of Clinical Epidemiology & Biostatistics, McMaster University, Canada; 3 Department of Health Promotion and Human Behavior, School of Public Health, Kyoto University Graduate School of Medicine, Japan; 4 Seattle Quality of Life Group, Department of Health Services, University of Washington, USA; 5 Faculty of Medicine, University of Calgary, Canada; 6 Department of Medical Psychology J3-220-1, Academic Medical Center, The Netherlands; 7 Iberoamerican Cochrane Centre, Institute of Biomedical Research (IIB Sant Pau), Spain; 8 Centre for Evidence-Based Chinese Medicine, Beijing University of Chinese Medicine, Department of Clinical Epidemiology & Biostatistics, McMaster University, China; 9 Department of Health Promotion and Human Behavior, Kyoto University Graduate School of Medicine/School of Public Health, Japan; 10 Department for Evidence-based Medicine and Clinical Epidemiology, Danube University Krems, Austria; 11 Hospital Civil de Guadalajara, Fray Antonio Alcalde, Mexico; 12 Nephrology Program, Humber River Regional Hospital; Division of Nephrology, University of Western Ontario, Canada Background: The ability to interpret and, consequently, include patient-reported outcomes (PROs) in the decision-making process along with other patient-important outcomes relies on the understanding of what improvement or reduction in the PRO represents a change that is important to patients. With such understanding, treatment effects can be understood in terms of their magnitude as large, moderate, small but still important, or negligible. The most common reference point for the interpretation of PROs is the minimal important difference (MID), which provides a measure of the smallest change in a PRO that patients would perceive as an important benefit or harm. Although MIDs are available for many PROs, users of these MIDs have no guidance on how to determine how credible these estimates are. Objective: To develop an instrument to measure the credibility of MIDs estimates studies. Methods: We are currently conducting a systematic survey that summarizes and appraises available methods to estimate anchor-based MIDs. We searched MEDLINE, EMBASE, PsycINFO, and CINAHL (1989 to present) to identify studies addressing methods to estimate anchor-based MIDs of PRO instruments or reporting empirical ascertainment of anchor-based MIDs. Teams of two reviewers will independently screen the citations identified, and extract relevant data. We will summarize the available methods and develop a new instrument addressing the credibility of empirically ascertained MIDs. We will user-test and pilot our Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. tool, and assess its inter-rater reliability. Discussion: Our new instrument will allow users to determine the extent to which the design and conduct of studies measuring MIDs is likely to have protected against misleading estimates. In addition, researchers interested in generating these estimates can learn about key aspects that need to be considered in MID study designs to minimize bias. We will present the methods and final draft of the instrument. LRO 1.5 Meta-analysis of mean differences in minimal important difference units: application with appropriate variance calculations Shrier I1 , Christensen R2 , Juhl C3 , Beyene J4 1 Centre for Clinical Epidemiology, Lady Davis Institute, McGill University, Canada; 2 Musculoskeletal Statistics Unit, The Parker Institute, Bispebjerg and Frederiksberg Hospital, Denmark; 3 Research Unit for Musculoskeletal Function and Physiotherapy, University of Southern Denmark, Denmark; 4 Department of Mathematics and Statistics, McMaster University, Canada Background: Practicing evidence-based medicine requires succinctly summarized data, preferably in a meta-analysis if the data are appropriate. For continuous outcomes using different measurement instruments, systematic reviewers may calculate the standardized mean difference (SMD), ratio of means (RoM), or a more recent method based on mean difference between groups expressed in minimal important differences (MD-mid) units, where the MID is considered a constant. Although standardization in MID units is easily interpretable clinically, considering MID as a constant imposes important limitations. Objectives: Our objective is to illustrate how considering the MID as a random variable, with a distribution, provides solutions to these limitations. Methods: We calculated the variance of MD-mid using the delta method. Using sensitivity analyses, we compare results using the two methods of calculating the MD-mid variance. Results: Considering the MD-mid as a random variable instead of a constant, (1) enables investigators to obtain estimates for questionnaires with no previous MID, and (2) makes underlying assumptions more transparent. Furthermore., considering MD-mid as a random variable allows investigators to avoid making the unrealistic assumptions that (1) the coefficient of variation for MID is independent of the measure, and (2) there is no correlation between the MID and mean difference. Using sensitivity analyses for different assumptions, we illustrate that the variance of MD-mid calculated when MID is considered a constant instead of a random variable can be under or over-estimated to significant degrees. We explore the effects on two different datasets, (1) data originally used to present the MD-mid, and (2) data from osteoarthritis studies using different pain scales and disability scales. Conclusions: Considering the MID as a random variable instead of a constant makes underlying assumptions more transparent and accounts more appropriately for the true variance. Cochrane Database Syst Rev Suppl 1–327 (2015) 33 LRO 1.6 Communicating the magnitude of treatment effects: a survey of clinicians in eight countries Johnston B1 , Alonso-Coello P2 , Friedrich J3 , Mustafa R4 , Tikkinen K5 , Neumann I6 , Vandvik P7 , Akl E8 , da Costa B9 , Adhikari N3 , Mas Dalmau G2 , Kosunen E10 , Mustonen J10 , Crawford M11 , Thabane L12 , Guyatt G12 1 The Hospital for Sick Children Research Institute, University of Toronto, Canada; 2 Iberoamerican Cochrane Center, Spain; 3 University of Toronto, Canada; 4 University of Missouri-Kansas City, USA; 5 University of Helsinki, Finland; 6 Universidad Catolica de Chile, Chile; 7 University of Oslo, Norway; 8 American University of Beirut, Lebanon; 9 University of Bern, Switzerland; 10 University of Tampere, Finland; 11 The Hospital for Sick Children Research Institute, Canada; 12 McMaster University, Canada Background: Meta-analyses of continuous outcomes typically provide enough information for decision makers to evaluate the extent to which chance can explain apparent Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. differences between interventions. The interpretation of the magnitude of these differences - from trivial to large can, however, be challenging. Objectives: To investigate clinicians’ understanding and perceptions of usefulness of six statistical formats for presenting continuous outcomes from meta-analyses: standardized mean difference (SMD), minimal important difference units (MID), mean difference (MD), ratio of means (RoM), relative risk (RR), and risk difference (RD). Methods: We invited 610 staff and trainees in family medicine and internal medicine programs in eight countries (Europe, North and South America, and the Middle East) to participate. Paper-based, self-administered questionnaires presented summary estimates of hypothetical interventions versus placebo for chronic pain. The estimates were either for a small or large effect for each of the six presentation approaches. Questions addressed participants’ understanding of the magnitude of treatment effects and their perception of usefulness of the format. We randomized participants to size of effect and order of presentation of the different formats. Results: 531 (87%) of clinicians responded (Table 1). Clinicians best understood RD, followed by the RR and RoM (Figure 1). Similarly, clinicians perceived Cochrane Database Syst Rev Suppl 1–327 (2015) 34 Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Cochrane Database Syst Rev Suppl 1–327 (2015) 35 dichotomous presentation of continuous outcomes (RD; RR) most useful (Figure 2). Compared to family medicine, internal medicine had better understanding and reported higher perceived usefulness of the various presentation formats (P value < 0.005; Figure 3). Conclusions: No format was very well understood or perceived as extremely useful. Clinicians best understood continuous outcomes when presented as dichotomies (relative and absolute risk differences) and perceived these presentations to be the most useful. Presenting results as SMD, the longest standing and most widely used approach, was poorly understood and perceived as least useful. Further efforts in medical education regarding understanding research results are necessary. Lunch rapid oral session 2 Priorisation LRO 2.1 The UK top 10 clinical research priorities in neuro-oncology Bulbeck H1 , Grant R2 , Oliver K3 , Day J4 , Macdonald L5 , Morley R6 1 Brainstrust, United Kingdom; 2 Department of Clinical Sciences, Western General Hospital, Edinburgh, United Kingdom; 3 International Brain Tumour Alliance, United Kingdom; 4 NHS Lothian, United Kingdom; 5 Cochrane Collaboration, United Kingdom; 6 James Lind Alliance, United Kingdom Background: The James Lind Alliance (JLA) supports the National Institute of Health Research (NIHR) to develop research evidence. Many clinical questions about brain and spinal cord tumours remain unanswered –therefore patients, carers and clinicians should work collaboratively to establish the most important research priorities. A Neuro-Oncology Priority Setting Partnership was established following JLA guidelines to identify the top 10 uncertainties in diagnosis, treatment and management of primary central nervous system (CNS) tumours. Objectives: The primary objective of the Neuro-Oncology PSP was to identify the top 10 treatment uncertainties in diagnosis and management of primary CNS tumours. A secondary objective was to highlight the necessity of a brain tumour community collaborative approach to priority setting in clinical research. Methods: Stakeholder representatives from a cross-section of the UK brain tumour community were recruited. The scope was clinical interventions for primary brain or spinal cord tumours. The first public survey was widely publicised to patient forums, professional organisations and charities. Questions were categorised and standardised into PICO (participants, interventions, comparisons, outcomes) format. A systematic literature search was performed to ensure the questions were genuine ’uncertainties’. Initial prioritisation by stakeholders was followed by a second public survey. A final workshop determined the top 10 questions using a modified Delphi and Nominal Group technique. Results: Over 600 individual Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. questions were generated from the initial survey, a patient forum and the UK Database of Uncertainties about the Effects of Treatment (DUETs). After analysis and selection, these reduced to 44 PICO questions. The second public survey reduced these further. The last 25 questions were discussed at a full-day stakeholders’ workshop. The final top 10 priorities were agreed. Conclusions: The brain tumour community has worked collaboratively to identify the top 10 priorities for clinical research. These will be promoted to governmental, neuroscience and charity funders throughout 2015/16. LRO 2.2 Setting priorities for Cochrane Reviews in health communication and participation Synnot A1 , O’Sullivan M2 , Hill S2 1 Consumers and Communication and National Trauma Research Institute, Australia; 2 Consumers and Communication Review Group, Australia Background: Increasingly, the research community, including Cochrane, is looking to include consumers, carers, health professionals and policy makers to work in partnership with researchers to set research priorities. The Cochrane Consumers and Communication Review Group (CC&CRG) is undertaking a research priority-setting project to identify new topics for Cochrane Reviews (CRs) within our scope. To our knowledge, this will be the first project of this nature to have such a broad scope encompassing many areas of communication and participation in health. Objectives: We have two objectives: (1) to identify Australian and international priority topics for systematic reviews in health communication and participation; and (2), to identify the top five Australian priorities for new CRs within the scope of the CC&CRG. Methods: Drawing upon existing priority-setting frameworks, we will undertake a multi-stage project involving an online survey, a face-to-face workshop and evidence mapping. We will involve key stakeholders (consumers, carers, health professionals, policy makers and researchers) from Australia and internationally, in every stage and the results will be widely disseminated. Results: A multidisciplinary steering group has been convened to oversee the project and will be responsible for defining the project scope, establishing criteria for identifying priorities and providing guidance over the course of the project. Strategic partnerships with local stakeholders are being sought to enhance dissemination and translation opportunities for new CR topics and related outcomes. Subsequent project stages (online survey, face-to-face workshop) will be completed by October 2015. Conclusions: Upon completion, we will have up to five new topics for CRs. In 2016, we will seek author teams to undertake these high-priority reviews. Novel consumer co-production models may be trialled as part of the review commissioning process. Cochrane Database Syst Rev Suppl 1–327 (2015) 36 LRO 2.3 LRO 2.4 Prevalence of insurance medicine related outcomes in update reviews of Cochrane’s 2015 priority list Use of framework synthesis to prioritise systematic review topics among multiple stakeholders: a descriptive case study Kunz R1 , Vogel N1 , deBoer W1 , von Allmen D1 1 Swiss Academy of Insurance Medicine, University Hospital Basel, Switzerland Brunton G1 , Llewellyn A2 , Sutcliffe K1 , Rees R1 , Oliver S1 , Caird J1 , Stokes G1 , Thomas J1 1 EPPI-Centre, UCL IOE London, United Kingdom; 2 Centre for Reviews & Dissemination, University of York, United Kingdom Background: Cochrane Insurance Medicine has recently registered as a Field with Cochrane, and went public in March 2015. Insurance medicine deals with health-related issues in the context of insurance coverage, in particular sick leave, short- and long-term disability, and interventions to facilitate return to work. While almost all medical disciplines benefit from outcome research and their synthesis in the form of systematic reviews, insurance medicine has missed out this development. Equally, many healthcare areas ignore the importance of outcomes related to sick leave and short-and long-term disability, and, thereby, neglected important consequences of health care to patients and society. Objectives: To explore the prevalence of insurance medicine related outcomes in updated reviews on Cochrane’s 2015 priority list (editorial-unit.cochrane.org/cochrane-priority-review-list2015-16). Methods: We used a cohort of systemic reviews (SR) from the Cochrane priority list for updates: Population: SRs on adults of working age with health conditions that commonly result in temporary or permanent inability to work (sick leave, short- and long-term disability). Interventions: Healthcare interventions that may impact on the course and/or duration of the disease. Outcomes: Duration of sick leave, short- and long-term disability and their prevention; return-to-work and other work-related disability outcomes as defined by Cochrane Occupational Safety and Health Group (osh.cochrane.org). We will report type and frequency of primary and secondary outcomes. Results: The priority list identified 144 of 297 review topics from 49 Cochrane Review Groups with an urgent need for update, 100 of which met our inclusion criteria. The analysis is currently underway. Results will be presented at the Colloquium. Conclusions: We anticipate that despite their importance for individuals and society, very few reviews (< 5%) will consider outcomes related to reduction of sick leave, short- or long-term disability, and successful return to work. Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Background: Systematic reviewers are challenged to mediate between an increasing amount of research and stakeholder priorities about the most important synthesis topics. Framework synthesis offers an innovative method to meet both needs. Objectives: To describe and reflect on framework synthesis as both a method of research synthesis and to scope and set research priorities for policy decisions, using a recent systematic review of extrahepatic conditions in hepatitis C. Methods: To familiarise researchers with the review topic, scoping searches were undertaken and an a priori framework of extrahepatic conditions selected. An information scientist searched for quantitative studies examining associations between hepatitis C and any extrahepatic condition. Included studies were indexed/coded into the framework (i.e. coded) by their main extrahepatic condition and any newly identified conditions added. Two reviewers screened and coded reports, establishing inter-rater reliability, then undertook lone assessments. To chart the data, results were ordered by publication frequency and presented to public advocacy groups and researchers, whose prioritised extrahepatic conditions were compared and added to the framework. To interpret the findings, topics most frequently reported and also identified by all advocacy groups were prioritised for further synthesis. Results: Searches identified 194 extrahepatic conditions. Framework synthesis structured stakeholder discussions to prioritise extrahepatic conditions and identified disparities between most frequently researched conditions and those identified as important by people experiencing the conditions. Differences in presentation of the framework during consultations may have influenced which topics each group prioritised. Conclusions: Framework synthesis integrated different perspectives, facilitating transparent research decisions. Use of an a priori framework enabled coding within short timelines. The differing roles and priorities of consulted stakeholders suggest that equal involvement is more likely, and response bias reduced, if all stakeholder groups are given consistent information. Cochrane Database Syst Rev Suppl 1–327 (2015) 37 LRO 2.5 Using cumulative meta-analysis methods for prioritizing Cochrane Review updates Spence G1 , Fanshawe T1 1 University of Oxford, United Kingdom Background: The regular updating of Cochrane Reviews is a demanding, ongoing process. To focus resources, reviews in which the pooled results are inconclusive or highly likely to be altered by new studies should be prioritized over those whose results are more stable, or even definitive. As such, quantitative methods to assess the strength of evidence within meta-analyses are desirable. To monitor the accumulation of evidence over time, cumulative meta-analysis can be undertaken in which the effect size estimate is updated sequentially as new results become available. Objectives: The aim of this study is to compare alternative statistical methods for assessing whether sufficient data have been accumulated at different points within cumulative meta-analyses, to determine whether a review update should be prioritized. Methods: Several methods to quantify the strength of evidence within a meta-analysis were applied to a database of the systematic reviews published or updated in the Cochrane Library between 2008 and 2012. Such methods–including the calculation of monitoring boundaries based on an ‘Optimal Information Size’ –have mostly been adapted from the interim analysis of clinical trials, in which the same issues of multiplicity from repeated testing and random fluctuations in effect size are important. Results: The methods investigated in this study were appraised by comparing their application to a number of cumulative meta-analyses. Differences in the results obtained from the various methods will be demonstrated with examples, and the suitability of the methods in different situations compared. Conclusions: We draw conclusions about the attributes and suitability of different statistical methods of assessing the strength of evidence within a meta-analysis. Furthermore, such methods can provide an objective, evidence-based approach to recommendations for systematic review updates. LRO 2.6 Focus on registered studies, adequately powered analyses and true data: filter unreliable evidence from Cochrane Reviews! Sydenham E1 1 Cochrane Injuries Group, United Kingdom Background: Information on the effectiveness and safety of health care interventions should be valid, precise, up-to-date, comprehensible and freely available. Currently none of these criteria is adequately satisfied. Many systematic reviews contain a biased sample of under-powered trials (some include fabricated trials) and reviews are often long and unreadable. Methods: In response to these concerns, the Cochrane Injuries Group developed a new policy (Roberts Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. 2015). In addition to following standard Cochrane methods, authors of new and updated reviews should: 1. Only include a trial if it was prospectively registered (unless the trial report was published before 2010). 2. Provide a sample size estimate showing how many participants need to be included in a meta-analysis for reliable results. Trial sequential methods should be used to explore all ’significant’ treatment effects obtained before reaching the required sample size. 3. The editorial team will conduct statistical checks on data included in a review. Authors of selected trials will be asked to provide the original trial data for checking. If authors decline to provide it, the trial will be removed from the review. Conclusions: We invite critical discussion of our policy and urge other review groups to consider adopting it. Reference: Roberts I, Ker K, Edwards P, Beecher D, Manno D, Sydenham E. The knowledge system underpinning health care is not fit for purpose and must change. BMJ 2015: 350: h2463. Lunch rapid oral session 3 Information retrieval LRO 3.1 An evidence-based approach to scoping reviews Khalil H1 , Peters M2 , Godfrey CM3 , McInerney P4 , Baldini Soares C5 , Parker D6 1 Monash University, School of Rural Health, Australia; 2 The University of Adelaide, The Joanna Briggs Institute, Australia; 3 Queen’s University, School of Nursing, canada; 4 University of Wits, South Africa; 5 University of São Paulo, Brazil; 6 The University of Queensland, Australia Background: The Joanna Briggs Institute (JBI) has as its central focus not only effectiveness, but also appropriateness, meaningfulness and feasibility of health practices and delivery methods. These questions are often answered by considering other forms of research evidence. Scoping reviews are used to assess the extent of a body of literature on a particular topic, and often to ensure that further research in that area is a beneficial addition to world knowledge. Objectives: To develop a methodology for scoping reviews based on the JBI focus on literature synthesis. Methods: A working group consisting of members of the Joanna Briggs collaborating organisations met to discuss the proposed framework for the methodology and develop a draft for the scoping review methodology based on the Arksey and O’Malley framework (2005) and that of Levac, Colquhoun and O’Brien (2010). This was followed by a workshop attended by other members of the organisations consisting of 30 international researchers to discuss the proposed methodology. Further refinement of the methodology was undertaken as a result of the feedback received from the workshop. Results: The development of the methodology focussed on five stages of the protocol and review development. These were: identifying the research question by clarifying and linking the purpose and research question; identifying the relevant studies using a three-step Cochrane Database Syst Rev Suppl 1–327 (2015) 38 literature search in order to balance feasibility with breadth and comprehensiveness; careful selection of the studies using a team approach; charting the data in a tabular and narrative format; and collating the results to identify the implications of the study findings for policy, practice, or research. Conclusions: Following the JBI trend of valuing different types of scientific literature in order to achieve better health care for all, a proposed methodology was developed based on the Arksey and O’Malley and the Levac, Colquhoun and O’Brien frameworks and was successfully trialled. LRO 3.2 Development of an objectively derived and validated search strategy to retrieve overviews of systematic reviews in MEDLINE Lunny C1 , McKenzie J1 , McDonald S1 , Lunny C2 1 Australasian Cochrane Group, Australia; 2 Australasian Cochrane Group, Monash University, Australia Background: There has been a steady increase in the number of overviews of systematic reviews published in the last 15 Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. years. The ability to dentify published overviews efficiently is of importance to users of evidence syntheses, as well as to researchers interested in investigating their methods. Locating overviews is difficult due to a lack of a validated search strategy and inconsistent terminology used to describe overviews. Objectives: To develop a validated search strategy to retrieve overviews of systematic reviews in MEDLINE. Methods: We derived a ’gold standard’ test set of overviews published between 1998 and 2011 from the references of two methods papers. Two population sets were used to identify discriminating terms; that is, terms that appear frequently in the test set, but infrequently in two population sets of references found in MEDLINE. We used a text mining package in the statistical program R to conduct a frequency analysis of terms appearing in the titles and abstracts. Terms that had a sensitivity of 10% or more in the test set and a corresponding low sensitivity (less than 2%) in the two population sets were selected for further testing. Candidate terms were combined in search strategies and tested in MEDLINE. Sensitivity and precision were used to evaluate filter performance. We evaluated the performance of the strategy against an externally derived set of overviews published between 2012 and 2014. Results: Two search strategies were Cochrane Database Syst Rev Suppl 1–327 (2015) 39 developed. The sensitivity-maximising strategy offers 93% sensitivity in retrieving overviews of reviews, with a low level of precision (4%). The sensitivity-and-precision-maximising strategy offers a less sensitive strategy (66%), but a higher precision (21%). Conclusions: We have developed two validated search strategies for locating overviews of systematic reviews using a text mining approach. Consistent language used to describe overviews of reviews would aid in their identification, as would a specific MEDLINE publication type. LRO 3.3 More than expert searchers? A case study for librarians exploring roles beyond databases Ayala AP1 , Lenton E1 1 Gerstein Science Information Centre, University of Toronto, Canada Background: At the University of Toronto (UofT), librarian involvement in a variety of reviews and grant proposals happens largely and primarily on an ad-hoc basis. The role of health sciences and medical librarians as expert searchers is well established and recognized, yet librarians can, and have, also been called upon to navigate other components of the research process (Dudden 2011). Involvement includes but is not limited to: project manager, research coordinator, and process/protocol expert. Librarian involvement is not only recommended, but has become a requirement for grants from key funding agencies, such as the Canadian Institutes of Health Research (CIHR), for a proposal to be considered complete and admissible (CIHR 2010). This is wonderful news, but are we prepared? Objectives: The objective of this study is to develop Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. a sustainable set of service standards in order to clarify the roles and tasks to be performed by librarians in research initiatives such as systematic reviews, but perhaps more importantly, scoping and realist reviews, as support requests for these newer, less familiar, methodologies have become increasingly common. Methods: We will gather and organize pre-established protocols for each type of research study and disseminate them widely. An environmental scan will be performed to learn more about how librarians at UofT are currently supporting research teams’ reviews. This knowledge can then be shared among our colleagues and incorporated into the content of our instruction portfolio. Results: The results of this initiative and anecdotal gathering are ongoing and will be presented at the Colloquium. Conclusions: Librarians offer significant contribution and insight beyond their searching skills; they can successfully act as project managers, process experts, and disseminators of protocols (communicators). Their role continues to evolve. Above all our goal remains to have an open and honest conversation on what librarians and information specialists’ experiences have been so far in these type of initiatives, what our roles are and what they should be. LRO 3.4 The impact of selective searching on the results of systematic reviews Hartling L1 , Featherstone R2 , Nuspl M2 , Shave K2 , Vandermeer B2 1 Child Health Field; University of Alberta, Canada; 2 Alberta Research Center for Health Evidence; University of Alberta, Canada Background: One of the hallmarks of a well-conducted systematic review (SR) is a comprehensive search. However, Cochrane Database Syst Rev Suppl 1–327 (2015) 40 empiric evidence is lacking on the impact of database selection on the results of SRs. Objectives: To examine the impact on the results of SRs based on selective searching. Methods: The sample was all child-relevant SRs with at least one meta-analysis from three Cochrane Review Groups (Acute Respiratory Infections (ARI), Infectious Diseases, Developmental Psychosocial and Learning Disorders; n = 129). We searched 13 databases to determine the proportion of relevant studies that were indexed in each source. We re-ran the primary meta-analysis with the studies identified in MEDLINE only and in MEDLINE plus each of the other databases. We calculated a ratio of the new point estimate to the original point estimate and a ratio of confidence interval widths. We determined how often results changed in statistical significance. Results: These results are based on the 57 SRs from ARI; complete results for all SRs will be available in October 2015. The average number of relevant studies was highest for MEDLINE (85%), EMBASE (80%), and BIOSIS (65%). From a total of 398 studies, 65 were not found in MEDLINE; 20 of the 65 were found in EMBASE and 13 in BIOSIS; other databases identified three or less. The mean ratio of point estimates for MEDLINE to the reference standard for dichotomous outcomes (n = 37) was 1.06. The smallest ratio was 1.03 for MEDLINE+BIOSIS. The mean ratio of confidence interval widths of MEDLINE to the reference standard was 1.12; the smallest ratio was 1.09 for MEDLINE+EMBASE. The mean ratio of standardized mean differences (n = 19) was 0.01 for MEDLINE alone, with the mean ratio of confidence interval 1.17 (smallest for MEDLINE+EMBASE, 1.15). For MEDLINE alone, five meta-analyses changed in significance. The fewest number of meta-analyses changing in significance (n = 1) was for MEDLINE+EMBASE. Conclusions: This study provides quantitative data regarding the impact on SR results of restricting searches to select databases. This information may be useful to increase efficiencies in the conduct of SRs. LRO 3.5 Will web search engines replace bibliographic databases? Review Taylor B1 , Best P2 , McQuilkin J1 1 Ulster University, United Kingdom; 2 Queen’s University Belfast, United Kingdom Background: The task of identifying relevant research is becoming more demanding, particularly for retrieving research on psychosocial as well as pharmacological interventions and for studies using diverse research designs suited to various stages of developing health and social care interventions (Medical Research Council 2008). Search methods must address precision (rejecting unwanted papers) as well as sensitivity (retrieving relevant papers). The advent of web search engines adds a new dimension to the usual searching of bibliographic databases. Objectives: To review papers reporting empirical studies that compare searching the world wide web with searching bibliographic databases to retrieve health and social care research. Methods: An ‘ideal’ formula involving 73 search terms (plus truncation variants) Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. in three concept groups on the review topic retrieved over 20,000 hits on just one database. This search was optimised in terms of precision to make it manageable whilst retaining maximum possible sensitivity by eliminating less precise terms in sequence. The search was run on six bibliographic databases and four web search engines to identify papers about empirical studies on searching databases and the world wide web for research on health and social care. Papers were excluded if they did not study the relative merits of databases and web search engines in terms of sensitivity or precision. Results: Results regarding sensitivity and precision of searching bibliographic databases and the world wide web will be summarised in tabular format. Conclusions regarding constructing search formulae will be summarised in narrative form with examples from papers and drawing on our own search process. Preliminary results suggest that: (1) Number Needed to Read is a readily-understood metric for precision; and (2) web search engines are not sufficiently sensitive or precise to replace bibliographic databases. Conclusions: Papers will be discussed in the broader theoretical context of ‘satisficing’ in relation to information-seeking behaviour (Prabha 2007; Simon 1956). Conclusions will be drawn for optimising precision of search strategies. References: 1. Medical Research Council (2008) Developing and Evaluating Complex Interventions, London: MRC Prabha C, Connaway LS, Olszewski L & Jenkins LR (2007) What is enough? Satisficing information needs. Journal of Documentation 63(1), 74–89. 2. Simon HA (1956) Rational choice and the structure of the environment. Psychological Review 63, 129–138. LRO 3.6 Googling for grey: using Google and Duckduckgo to find grey literature Hagstrom C1 , Kendall S2 , Cunningham H1 1 University of Toronto, Canada; 2 Mt. Sinai Hospital Library, Canada Background: Using search engines such as Google or Duckduckgo for literature searches can result in literally millions of citations. Results depend on various elements, such as the search engine’s page ranking, or the user’s habitual search patterns. Objectives: Google and Duckduckgo can be effectively used to retrieve government or other institutional reports and other hard-to-find grey literature. Methods: We will demonstrate how Google and Duckduckgo function. We will compare the two search engines and will give examples of how they can be searched to eliminate many irrelevant results. Results: Different types of grey literature can be found using Google and Duckduckgo. Conclusions: Using search engines such as Google and Duckduckgo need not be as onerous as one might think when looking for grey literature. Cochrane Database Syst Rev Suppl 1–327 (2015) 41 Lunch rapid oral session 4 Decision-making: End users and consumers LRO 4.1 Are online learning resources effective to build capacity for evidence-informed decision making? Dobbins M1 , Yost J1 1 McMaster University, Canada Background: Public health professionals are expected to consider the best available research evidence when making practice decisions. However, they encounter challenges such as time constraints, limited resources and skills to find, appraise, interpret and apply evidence. Objectives: To evaluate the usage and change in learning outcomes associated with use of online learning resources for evidence-informed decision making (EIDM). Methods: The National Collaborating Centre for Methods and Tools worked Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. with Web developers and online learning specialists to create the Learning Centre. As a management system, the Learning Centre offers multiple online learning resources for which the content, practical examples and assessments have been developed in collaboration with public health professionals. Quantitative and qualitative evaluation data are continuously collected before and after use of the online learning resources. Results: Evaluation data from June 2012, when the Learning Centre was launched, to May and September 2015, will be presented. As of March 2015, the resources in the Learning Centre have been used by more than 3000 public health professionals. Quantitative data demonstrate significant increases in knowledge and self-efficacy for EIDM following use of the resources. Quantitative and qualitative data indicate learners are satisfied with the online format of the Learning Centre; users return to the resources multiple times, suggesting they value the resources as references to support EIDM capacity. Conclusions: Evaluation of this project validates that the online learning resources offered by the National Collaborating Centre for Methods and Tools are innovative, accessible and effective in supporting EIDM capacity development among public health professionals. Cochrane Database Syst Rev Suppl 1–327 (2015) 42 LRO 4.2 LRO 4.3 Outcomes of the use of targeted online consumer health information in primary care: a participatory systematic mixed-studies review Identifying patient-important practical issues for shared decision making Pluye P1 , El Sherif R1 , Granikov V1 , Tang DL1 , Bujold M1 , Hong QN1 , Galvao C2 , Repchinsky CA3 , Hutsul J3 , Dunikowski L4 , Frati F1 , Shoet L5 , Vineberg L6 , Bartlett G1 , Grad R1 , Shulha M1 , Kloda L1 , Burnand B7 , Desroches S8 , Legare F8 , Millerand F9 , Rihoux B10 , Vedel I1 1 McGill University, Canada; 2 Universidade de São Paulo, Brazil; 3 Canadian Pharmacists Association, Canada; 4 College of Family Physicians of Canada, Canada; 5 Center for Literacy, Canada; 6 Patient Representative, Canada; 7 Institut Universitaire de Médecine Sociale et Préventive, Switzerland; 8 Universite Laval, Canada; 9 Universite du Quebec a Montreal, Canada; 10 Universite Catholique de Louvain, Belgium Background: Consumers regularly seek online health information. Systematic reviews in public health and oncology, and on decision aids, suggest the use of targeted information improves knowledge, participation in health care and health. However, little is known about the outcomes associated with information-use in primary care (outside shared decision-making aids). Thus, for primary healthcare patients, what are the outcomes associated with the use of targeted (vs non-targeted) online consumer health information? Objective: Identify types of patient health outcomes and related conditions (context, information-seeking, use). Methods: - Participatory systematic mixed studies review (integrating qualitative and quantitative evidence). - Eligibility criteria: qualitative, quantitative, or mixed-methods study; English or French; 1990 to 2014; community-based primary health care; online consumer health information-use. - Information sources: MEDLINE, Embase, PsycINFO, CINAHL, LISA, and grey literature. - Search strategy built by four specialized librarians. - Selection (DistillerSR) and critical appraisal (Mixed Methods Appraisal Tool) of included studies by two independent researchers. - Data extraction and synthesis: Deductive-inductive qualitative thematic data analysis by four researchers, and disambiguation of themes. Results: Flow diagram (Fig. 1): 50 included studies with diverse designs, four concern targeted information (one non-randomized, two quantitative descriptive, and one qualitative). Inter-rater reliability scores (kappa) indicate substantial agreement between researchers for selection (0.62) and appraisal (0.67). The qualitative synthesis suggests key conditions and negative/positive outcomes (e.g. decreased/increased worries). Conclusions: Results will allow us to explore ‘condition-outcome’ configurations (content and Boolean analysis). Using a mixed-methods research design, this is the first systematic review to explore outcomes of targeted online consumer health information-use in primary care. This will lead to design a tool to help patients for finding and using (e.g. with partners or clinicians) information such as decision aids. Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Fog Heen A1 , Montori VM2 , Lloyd A3 , Vandvik PO1 , Agoritsas T4 1 Department of Medicine, Innlandet Hospital Trust, Gjøvik, Norway and Institute for Health and Society, Faculty of Medicine, University of Oslo, Norway; 2 Knowledge and Evaluation Research Unit, Mayo Clinic, Rochester, Minnesota, USA; 3 Institute of Primary Care and Public Health, Cardiff University School of Medicine, United Kingdom; 4 Department of Clinical Epidemiology and Biostatistics, McMaster University, Canada Background: To help patients and clinicians engage in collaborative deliberation, tools for shared decision making (SDM) usually include evidence on the nature and likelihood of patient-important outcomes associated with management alternatives. However, patients face practical issues when implementing the chosen option in their daily life. Evidence summaries typically include the former but overlook the latter. Objectives: To map and categorize patient-important practical issues related to treatment options, and develop a generic framework for their incorporation into the creation of SDM tools. Methods: We systematically reviewed and mapped two main sources (HealthTalk.org registry and Option Grids) that included a large sample of conditions and applied a rigorous methodology to identifying patient experience and questions. We systematically screened every topic, abstracting all practical issues relevant to management options. Two independent reviewers grouped the issues in themes, compared results, resolved discrepancies and arrived at final a set of themes. We tested the applicability of this framework in identifying specific practical issues when creating 14 SDM tools using the SHARE-IT method. Results: Thematic analysis identified 15 categories from 967 issues across 297 topics in HealthTalk.org and 29 Option Grids: five were care-related (medication routine, tests and visits, procedure and device, recovery and adaptation, co-ordination of care); five related to daily life (food and drinks, exercise and activities, social life and relationships, work and education, travel and driving); and five related to miscellaneous issues (adverse effects, interactions and antidote, physical well-being, emotional well-being, pregnancy and nursing, costs and access). In creation of 14 SDM tools on antithrombotic therapy and chemotherapy, all categories were applicable and used. Conclusions: We identified 15 overarching themes of patient-important practical issues. This generic framework can complement trustworthy evidence summaries from systematic reviews and help thus create tools that are more helpful in discussing with patients issues that matter to them. Cochrane Database Syst Rev Suppl 1–327 (2015) 43 LRO 4.4 LRO 4.5 Combining network meta-analysis results with patient preferences to facilitate shared decision making Sources of information and attitudes towards evidence-based healthcare: a survey of Swiss physicians Naci H1 , van Valkenhoef G2 , Higgins J3 , Fleurence R4 , Ades T3 1 London School of Economics and Political Science, United Kingdom; 2 University of Groningen, The Netherlands; 3 University of Bristol, United Kingdom; 4 Patient Centered Outcomes Reserach Institute, USA Burnand B1 , Bengough T2 , Barry A3 , Pidoux V3 , Amiguet M3 , Bovet E4 1 Cochrane Switzerland and Institute of Social and Preventive Medicine, Lausanne University Hospital, Switzerland; 2 Austrian Federal Institute of Health Care, Austria; 3 Institute of Social and Preventive Medicine, Lausanne University Hospital, Switzerland; 4 Haute Ecole de Santé Vaud and Institute of Social and Preventive Medicine, Lausanne University Hospital, Switzerland Background: Even in cases when comparative evidence about drugs exists, prescribers and patients often struggle to weigh the relative benefits and harms of multiple options. One complexity of drug therapy is the difficulty in making trade-offs between the benefits and harms of two or more alternatives. Objectives: To formalize the incorporation of patient preferences into treatment selection decisions using statins as a case study. Methods: We combined network meta-analysis and multi-criteria decision analysis. Specifically, using a systematic review and network meta-analysis of randomized trials of statins, we calculated absolute risks of all-cause mortality, coronary events, cerebrovascular events, discontinuations due to adverse events, myalgia, transaminase elevations and creatine kinase elevations associated with each statin. We then applied a structured benefit risk model that allows evidence on multiple outcomes to be combined using qualitative preference statements. When combining the evidence on multiple outcomes, we adopted simple preference statements about the relative importance of different outcomes and considered the effect of statins on preventing mortality to be more important than either coronary or cerebrovascular events, which were in turn more important than any one of the harm outcomes. Results: Fluvastatin has a considerable probability of both being the best (41%) and worst (12%) statin (based on the combination of benefits and harms), highlighting the uncertainty in its evidence base. Both simvastatin and atorvastatin have a high probability of better rank, with a negligible probability of ranking worst: atorvastatin and simvastatin have the most favorable benefit and harm profiles. Conclusions: In the future, summaries of clinical evidence obtained from network meta-analyses can be combined with patient preferences, and considered alongside the knowledge and clinical expertise of prescribers when making treatment selection decisions. Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Background: Barriers to knowledge translation generate delays to adopt new effective interventions, persistence of obsolete treatments, overuse of ineffective interventions and underuse of effective care. Objectives: To examine current knowledge translation practices of Swiss physicians. Methods: We surveyed French- and German-speaking private practice physicians. Semi-structured open-ended interviews and focus groups targeting family physicians, psychiatrists, cardiologists, endocrinologists, and orthopaedic surgeons in Switzerland, and the existing literature, allowed us to draft an interview questionnaire for the computer-assisted web interview (CAWI) of a large sample of these groups of physicians (summer 2014). We targeted sources of and barriers to information, and adherence to evidence-based medicine (EBM) principles. We computed a score of adherence to EBM based on five items and used uni- and multivariate analyses. Results: Twenty-nine physicians (21 men) participated in the qualitative study and 985 in CAWI (15% participation, similar among the three Swiss language regions and physicians’ specialty). CAWI indicated that the main sources of information of family physicians included congresses (41%), practice guidelines (38%), colleagues (37%) and specialists (33%), systematic reviews (28%), knowledge syntheses (24%) and quality circles (21%). Perceived usefulness of EBM in daily practice was considered very high by 20%, high by 56%, low by 20% and very low by 4% of physicians. Adherence to EBM decreased slowly with age (-2.1 percentage-points/10 years), and was lower in endocrinologists, family physicians, orthopaedic surgeons and psychiatrists ( -7.0, -8.4, -12.8, -13.8 percentage-points, respectively) compared to cardiologists. Conclusion: We observed that physicians working in Switzerland use various sources of information to update their knowledge, but that no single source dominates. Evidence-based sources are often used and three-quarters of physicians adhere to its principles. The optimal source, type and format of information for busy physicians are uncertain. Cochrane Database Syst Rev Suppl 1–327 (2015) 44 LRO 4.6 Evidence synthesis activities of a hospital evidence-based practice center and impact on hospital decision making Jayakumar K1 , Lavenberg J1 , Mitchell M1 , Doshi J1 , Leas B1 , Goldmann D1 , Williams K1 , Brennan P1 , Umscheid C1 1 University of Pennsylvania, USA Background: Hospital evidence-based practice centers (EPCs) seek to implement evidence into local policy and practice, but their impact on institutional decision making is unclear. Objectives: To assess the evidence synthesis activities and impact of the University of Pennsylvania Health System’s Center for Evidence-based Practice (CEP), a hospital EPC. Methods: Descriptive analysis of CEP’s internal database of rapid systematic reviews since CEP’s inception (July 2006 to June 2014), and survey of review requestors from CEP’s last four fiscal years. Descriptive analysis examined requestor and report characteristics; a questionnaire examined report usability and impact, and requestor satisfaction (higher scores on five-point Likert scales reflected greater agreement). Results: CEP has completed 249 reviews since inception. The most common requestors were clinical departments (29%, n = 72), chief medical officers (19%, n = 47), and purchasing committees (14%, n = 35). The most common technologies reviewed were drugs (24%, n = 60), devices (19%, n = 48), tests (12%, n = 31) and care processes (12%, n = 31). The mean report completion time was 70 days. Thirty reports (12%) informed computerized decision support interventions. More than half of reports (56%, n = 139) were completed in the last four fiscal years for 65 requestors. Forty-six of the 64 eligible survey participants responded (72%). Requestors were satisfied with the report (mean = 4.4), and agreed it was delivered within the expected timeframe (mean = 4.4), informed their final decision (mean = 4.1), and that their final decision was consistent with report findings (mean = 4.0). Conclusions: Our findings suggest hospital EPCs can efficiently synthesize and disseminate evidence addressing a range of clinical topics for diverse stakeholders, and can influence decision making. Hospital EPCs may be an effective infrastructure paradigm for promoting evidence-based decision making. Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Oral session 4 Qualitative evidence O 4.1 Using purposive sampling within qualitative evidence synthesis: experiences from a Cochrane Review of parents’ views and experiences of vaccination communication Ames H1 , Glenton C2 , Lewin S3 , COMMVAC Project 21 1 The Norwegian Knowledge Centre for the Health Services, Norway; 2 The Norwegian Knowledge Centre for the Health Services and Cochrane Norway, Norway; 3 The Norwegian Knowledge Centre for the Health Services and South African Medical Research Council, Norway Background: Reviews of qualitative research, or qualitative evidence synthesis (QES), are increasingly common and are now being done by Cochrane. Identifying and including all possible studies is a goal for reviews of intervention effectiveness, but is not necessarily the best approach for QES. Too much data due to a large number of studies may threaten the quality of the analysis. Using purposive sampling of included studies is one way of reducing the number of articles to be analysed. Objectives: To describe an approach to purposively sampling included studies in a QES. Methods: We searched electronic databases to identify studies that met the following inclusion criteria: (1) qualitative methods of data collection and analysis; (2) parents or carers as participants; and (3) a focus on views and experiences of vaccination information. For studies that met the inclusion criteria, we extracted information regarding study country and focus and assessed data ’richness’ on a 1-5 scale. Studies meeting the inclusion criteria were sampled using a three-step purposive process: (1) We sampled any study from a low- or middle-income country (LMIC) to help ensure geographic spread, as the majority of studies on the topic were from high-income settings; (2) We sampled all studies that scored 3 or more on the data richness scale; (3) We reviewed the abstracts of the remaining studies and sampled those that answered the review question most closely and in the most detail. Results: Sixty studies met our inclusion criteria. We sampled seven studies from LMIC settings; 14 studies that scored more than 3 on the data richness scale; and five studies that answered the review question best, leaving 26 studies for data analysis. Conclusions: This systematic, three-step approach may prove useful to other researchers attempting to synthesise qualitative literature from a large number of studies. Strengths included that it allowed us to achieve a geographic spread of articles together with rich data that closely answered the review question. The weakness is that we may have overlooked articles that did not meet our sampling criteria, but would have contributed to the synthesis. Cochrane Database Syst Rev Suppl 1–327 (2015) 45 O 4.2 Sensitivity analysis in systematic reviews of qualitative evidence: an example from a mixed-methods systematic review Langer L1 , Stewart R1 , Winters N2 1 University of Johannesburg, South Africa; 2 University of Oxford, United Kingdom Background: The systematic synthesis of qualitative evidence is a recent development of systematic review methodology. Systematic reviews of qualitative evidence need to subscribe to the same methodological rigour and transparency that characterises reviews of quantitative evidence. As a result, reviewers should assess the sensitivity of the findings of any qualitative evidence synthesis to a priori defined variables. Objectives: To report on the methodological insights gained from a sensitivity analysis conducted as part of a qualitative evidence synthesis in a mixed-methods systematic review. Methods: We applied thematic synthesis to synthesise the findings from qualitative research in a mixed-methods systematic review. Prior to extracting themes from the primary research evidence, each study was subjected to a critical appraisal assessing the reliability of the study’s findings. After configuring findings in the thematic synthesis, the results from the synthesis were subjected to a sensitivity analysis. We assessed whether the synthesis results were sensitive Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. critical appraisal tool; (2) the applied qualitative research methodologies; and (3) the inclusion of individual studies yielding a larger than average number of themes. Results: We found that the results of the synthesis were sensitive to the inclusion of individual studies that yielded a larger than average number of themes. We also established that the exclusion of studies after the application of the critical appraisal tool would have generated a number of themes not identified in the included evidence. Conclusions: To ensure the rigour and transparency of qualitative evidence syntheses, both a critical appraisal of the qualitative primary evidence, as well as a sensitivity analysis of the synthesis results, using a priori defined variables are required. O 4.3 Unpackaging context dependence in systematic reviews of complex interventions: a methodological approach for managing double-layered complexity Parrott J1 , Handu D2 , Benson-Davies S2 1 Rutgers University, USA; 2 Academy of Nutrition and Dietetics, USA Background: Measures of effect size in systematic reviews of complex interventions, like pediatric weight management, are notoriously difficult to estimate meaningfully - in part due Cochrane Database Syst Rev Suppl 1–327 (2015) 46 to to: (1) the exclusion of studies after the application of the the highly contextually dependent nature of multicomponent interventions. In short, complex treatment configurations are compounded by variations in method of delivery, length of treatment, setting, etc. This is complexity on top of complexity. Objectives: To devise an analytic approach for managing the layered complexity of complex interventions and understanding how different configurations of study/treatment context characteristics could affect outcomes. Methods: Data from 73 controlled clinical trials of pediatric weight management interventions published since 2005 were extracted into an online platform using a detailed extraction form designed to capture 30 different treatment and intervention context characteristics. Multiple components analysis was used to construct a reduced dimension adjacency matrix, which was then used to derive theoretically-informed groups of similar treatment ’mixes’. Characteristic by group Chi-square analyses provided description of the identified groups. Random-effects meta-analysis was then used to estimate treatment effect differences between groups. As heterogeneity was still high, and based on the theoretical position that configurations of context characteristics (rather than single characteristics) affect treatment results, crisp set qualitative comparative analysis (QCA) was used to identify discrete ’paths’ to positive or negative weight status outcomes at six and 12 months post-treatment. Results: The analysis identified a limited number of treatment/context configurations that were consistently associated with both positive and negative weight status outcomes (Figure 1). Conclusions: A combination of analytic procedures may be used to manage and gain a deeper understanding of heterogeneity in context dependent multicomponent interventions. This mixed approach has great potential for use in program design by practitioners to identify conditions where particular treatment mixes are likely to lead to positive or negative outcomes. O 4.4 Linking Cochrane qualitative evidence syntheses with effectiveness reviews: experiences from the development of an EPOC qualitative evidence synthesis Munabi - Babigumira S1 , Glenton C1 , Lewin S2 , Fretheim A3 , Nabudere H4 1 Global Health Unit, Norwegian Knowledge Centre for the Health Services, Norway; 2 Global Health Unit, Norwegian Knowledge Centre for the Health Services; Health Systems Research Unit, Medical Research Council of South Africa, Tygerberg, Norway; 3 Global Health Unit, Norwegian Knowledge Centre for the Health Services; Institute of Health and Society, University of Oslo, Norway; 4 College of Health Sciences, Makerere University, Uganda Background: Qualitative evidence syntheses published in the Cochrane Library provide additional evidence to inform or supplement effectiveness reviews. However, the methods Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. used to link the findings of qualitative evidence syntheses and effectiveness reviews are not yet established, and tend to vary across review teams. Objectives: To discuss the methods used, and our experiences when supplementing evidence from the review ‘The effectiveness of policies promoting facility-based deliveries in reducing maternal and infant morbidity and mortality in low- and middle-income countries’ with the qualitative evidence synthesis: ‘Factors that influence the provision of intrapartum and postnatal care by skilled birth attendants in low-and middle-income countries’. Methods: When synthesizing qualitative evidence, we utilise a logic model that represents the theories and assumptions about the links between various inputs and activities through which skilled attendance at birth influences outcomes. Reviewers then map the review findings to the logical model; adapt or elaborate the model as needed; and subsequently assess how the findings relate to the results of the effectiveness review. In particular, we seek to provide explanations of how the context or other relevant findings influence the outcomes reported in the effectiveness review. In addition to presenting the methods we used, we will share our experiences from the process of linking the two reviews. Results: Close collaboration between the two review teams, facilitated by face-to-face contact, has contributed to the development of both reviews. This collaboration has enabled us to identify areas where the qualitative evidence synthesis can explore and explain the findings from the effectiveness review. Both reviews are in progress and additional results will be available at the time of the Colloquium. Conclusions: Methods for linking qualitative evidence synthesis reviews to effectiveness reviews are yet to be established. Close collaboration of review teams facilitates this process. Data from qualitative evidence synthesis provides useful additional information about how and why interventions work. Oral session 5 GRADE O 5.1 Defining and evaluating quality of evidence in ’Summary of findings’ tables Hultcranz M1 , Schunemann H2 , Guyatt G2 1 Swedish Council on Health Technology, Sweden; University, Canada 2 McMaster Background: Cochrane has adopted the GRADE (Grading of Recommendations Assessment, Development and Evaluation) approach to categorizing quality of evidence - most prominently seen in ’Summary of ’indings’ tables –defined as reflecting the extent of confidence that the estimates of effects are correct. This definition leaves ambiguity: does it refer to confidence in point estimates, or confidence in the range of possible estimates? Objective: To clarify the definition and approaches to judging Cochrane Database Syst Rev Suppl 1–327 (2015) 47 evidence quality. Method: Discussions, workshops, iterative refinement of ideas, and discussion at a GRADE Working Group meeting. Results: Based on this process, quality of evidence may be better considered as the confidence or certainty that a true effect lies on one side of a specified threshold, or within a chosen range. This interpretation raises the challenge of defining the threshold or range for making quality ratings. We developed five possible approaches for making this judgment. For guidelines, what we call a fully contextualized approach requires simultaneously considering all critical outcomes and their relative value. We defined four less contextualized approaches more appropriate for systematic reviews. We applied all five approaches to the outcome of myocardial infarction in a review of dual antiplatelet therapy (DAPT) versus single antiplatelet therapy (SAPT) in patients who have undergone stenting of their coronary arteries (Table). The four approaches most appropriate for systematic reviews involve judging certainty that the true effect lies within the 95% confidence interval; that the effect lies within some range other than the 95% confidence interval; that the effect is something other than no effect at all; and that the effect lies within ranges of what we might consider small, moderate, or large. Conclusions: The proposed approach is now under consideration by the GRADE Working Group. If adopted by GRADE, and subsequently by Cochrane, it will provide a useful clarification of how Cochrane authors can make quality of evidence judgments. piloting of the tool on qualitative evidence syntheses. Results: CERQual bases assessments of confidence on four components: 1. the methodological limitations of the individual studies contributing to a review finding, assessed using a quality-assessment tool for qualitative studies; 2. the coherence of each review finding, assessed by looking at the extent to which a review finding is based on data that are similar within and across multiple individual studies and/or incorporates convincing explanations for any variations across individual studies; 3. the relevance of a review finding, determined by the extent to which the primary studies supporting a review finding are applicable to the context specified in the review question; and 4. the adequacy of data supporting a review finding, assessed by an overall determination of the degree of richness or scope of the evidence and quantity of data supporting a review finding. After assessing each component, an overall judgement of the confidence in each review finding is made. Confidence is assessed as high, moderate, low, or very low. Conclusions: GRADE-CERQual provides a transparent method for assessing the confidence of evidence from reviews of qualitative research. Like other GRADE tools, it may facilitate the use of these findings alongside reviews of effects, in Health Technology Assessments and in guideline development processes. O 5.3 O 5.2 Assessing how much confidence to place in findings from qualitative evidence syntheses: a new version of the GRADE-CERQual tool Glenton C1 , Lewin S2 , Booth A3 , Noyes J4 , Garside R5 , Tuncalp O6 , Carlsen B7 , Bohren M6 , Wainwright M8 , Rashidian A9 , Colvin C8 , Munthe-Kaas H1 1 Norwegian Knowledge Centre for the Health Services, Norway; 2 Norwegian Knowledge Centre for the Health Services; Medical Research Council of South Africa, Norway; 3 University of Sheffield, United Kingdom; 4 Bangor University, United Kingdom; 5 University of Essex Medical School, United Kingdom; 6 World Health Organization, Switzerland; 7 Uni Research Rokkan Centre, Norway; 8 University of Cape Town, South Africa; 9 Tehran University of Medical Sciences, Iran Background: Systematic reviews of qualitative studies (qualitative evidence syntheses) are increasingly used to bring together findings from qualitative studies. In order to use the synthesised findings to inform decisions we need methods to assess how much confidence to place in these findings. Objectives: To describe the latest version of a Grading of Recommendations Assessment, Development and Evaluation (GRADE) tool for assessing how much confidence to place in findings from qualitative evidence syntheses. Methods: GRADE’s Confidence of the Evidence from Reviews of Qualitative research (CERQual) tool was developed through review of existing tools; working group discussions; and Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Using the GRADE approach to network meta-analysis Guyatt G1 , Murad H2 , Puhan M3 1 McMaster University, Canada; 2 Mayo Clinic, USA; 3 Institute of Social and Preventive Medicine, Zurich, Switzerland Background: Until recently, network meta-analyses (NMA) failed to address the certainty associated with different paired comparisons in a network. The GRADE working group has developed an approach to NMA that involves rating the certainty in estimates for each paired comparison. Objective: To illustrate GRADE’s approach to NMA. Method: The GRADE approach was developed through discussion and iterative refinement; we have applied the approach in a number of NMAs. The approach involves four steps: (1) present direct and indirect treatment estimates for each paired comparison; (2) rate the certainty of each direct and indirect estimate; (3) present the NMA estimate for each comparison; (4) rate the certainty of each NMA estimate. Results: Rating the certainty of each direct estimate is straightforward using established GRADE methodology. Rating the certainty for indirect estimates involves identifying, for each paired comparison, the 1st order loop (those loops that involve only a single additional intervention linked to the two treatments/nodes of interest) that contributes most to the indirect estimate; rating the certainty for the two direct comparisons contributing to the indirect comparison; and choosing the lower of the two ratings. If direct and indirect estimates are consistent, the higher rating of the two represents the overall certainty of the Cochrane Database Syst Rev Suppl 1–327 (2015) 48 network estimate. If they are inconsistent, GRADE suggests using the higher of the two, rather than the network estimate, as the best estimate of effect. Examples illustrate the utility of the approach. For instance, in a NMA addressing optimal fluid administration in septic patients, high certainty evidence was available for the comparison of starch vs crystalloid, but only very low certainty for gelatin vs crystalloid. A challenge to full application of the GRADE approach is difficulty in generating separate indirect effect estimates in NMAs with large numbers of comparisons. Conclusions: Since clinicians need to take into account not only best estimates, but also certainty, in making their management decisions, the GRADE approach enhances the utility and interpretation of NMA. O 5.4 Research synthesis to help going from evidence to decision: a practical application of the GRADE/DECIDE Evidence to Decision Framework Parmelli E1 , Amato L1 , Vecchi S1 , Minozzi S1 , Davoli M1 1 Department of Epidemiology, Lazio Regional Health Service, Italy Background: Healthcare systems are offered with a wide range of technologies and services, but they have to cope with decreasing resources and the uncertainty about what is effective and more appropriate. Making decisions about healthcare interventions is complex. Decisions should be informed by the best available evidence, which, in order to be comprehensive needs to take into account all the relevant aspects (e.g. efficacy, safety, equity, costs), and gathered within a limited time period. DECIDE (www.decide-collaboration.eu) is a project funded by the European Community that, using the GRADE (Grading of Recommendations Assessment, Development and Evaluation) methodology (www.gradeworkinggroup.org), aims at implementing strategies to enhance dissemination and communication of scientific evidence to support on-time evidence-based decision making. Objectives: To prepare an informative and exhaustive evidence synthesis to help a Regional Commission taking a coverage decision using the GRADE/DECIDE Evidence to Decision (EtD) Framework (see attached file). Methods: We performed a systematic review of the available evidence about the efficacy and safety of trans-aortic valve implantation (TAVI) for aortic stenosis in high risk patients. We also searched for information about economic evaluations, patients’ preferences, feasibility and equity. We used the EtD Framework to present the information to the members of the Regional Commission. Results: We identified 4072 records: four Health Technology Assessment (HTA) reports, 13 systematic reviews (SR), three randomised controlled trials and 15 observational studies were eligible for our topic (the single studies were not included in the SRs). We prepared two different EtD Frameworks for the two questions of interest: (1) TAVI versus surgery and (2) TAVI versus medical therapy. Conclusions: The relevance of this experience lies in the opportunity to pilot test the GRADE/DECIDE EtD Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Framework for real world coverage decisions. Feedbacks about its usefulness and usability will be collected from the members of the Regional Commission. Oral session 6 Disseminating evidence O 6.1 Lessons learnt from a planned, targeted dissemination of a systematic review on a controversial topic in the field of nutrition Naude CE1 , Schoonees A1 , Young T2 , Senekal M3 , Garner P4 , Volmink J2 1 Centre for Evidence-Based Health Care, Stellenbosch University, South Africa; 2 Centre for Evidence-Based Health Care, Stellenbosch University; South African Cochrane Centre, South African Medical Research Council, South Africa; 3 Division of Human Nutrition, University of Cape Town, South Africa; 4 Effective Health Care Research Consortium, Liverpool School of Tropical Medicine, United Kingdom Background: Low carbohydrate diets for weight loss and cardiovascular benefits have generated controversy between advocates, media and health professionals. At the request of the Heart and Stroke Foundation South Africa, we conducted a systematic review and actively disseminated its findings to inform ongoing debates. Objectives: To describe the stakeholder-specific dissemination strategy for a systematic review and assess its ‘reach’ and reflect on lessons learnt. Methods: We planned, developed and implemented a stakeholder-specific dissemination strategy, including identifying priority stakeholders, communication objectives and channels; tailoring approaches for stakeholders; developing appropriate products; and interacting with key role players. ‘Reach’, defined here as exposure to information, was measured by capturing all dissemination, including media monitoring and non-traditional metrics. Results: Stakeholders included scientists, health professionals, media, public and policymakers. Implementation resulted in wide ‘reach’, including 14 scientific presentations, two national television and five radio interviews, broad social media coverage, Altmetric score in the 99th percentile and more than 75 print and digital media contributions. The ‘reach’ was likely positively influenced by the controversy and public interest, as well as by interaction between the researchers and key stakeholders. The ‘reach’ may have increased stakeholders’ motivation and ability to use the evidence to inform decision-making. There is value in responding to stakeholders, producing a timely, rigorous review, disseminating it via a planned, targeted strategy, and being prepared and responsive to media requests. Conclusions: Planning and implementation of a targeted dissemination strategy for a systematic review is likely to increase the ‘reach’ and possibly use of rigorous evidence. Part Cochrane Database Syst Rev Suppl 1–327 (2015) 49 of the success related to the timeliness of completion. Review teams should be mindful that detailed, opportune planning and preparation of a dissemination strategy is essential, especially for controversial, media-hot topics. O 6.2 Cochrane 2.0: Tweeting and blogging to disseminate child health evidence Hartling L1 , Hamm M2 , Newton A3 , Fernandes R4 , Featherstone R5 , Thomson D1 1 Child Health Field, University of Alberta, Canada; 2 Alberta Research Center for Health Evidence, University of Alberta, Canada; 3 Department of Pediatrics, University of Alberta, Canada; 4 Child Health Field; Cochrane Portugal; Department of Pediatrics and Clinical Pharmacology Unit, Lisbon Academic Medical Centre, Canada; 5 Alberta Research Center for Health Evidence; University of Alberta, Canada Background: Healthcare providers desire ready access to reliable synthesized information to support point-of-care decision-making. Virtual communities, facilitated by the adoption of social media tools such as Facebook, Twitter, and YouTube, are increasingly used for knowledge mobilization, bridging the gap between knowledge generation/synthesis and knowledge implementation. Objectives: To implement and evaluate a structured social media strategy to disseminate high quality, child health evidence from Cochrane systematic reviews and overviews to healthcare providers caring for children. Methods: The Child Health Field’s social media strategy has three components: daily tweets, weekly blog posts, and a monthly Twitter journal club. Each tweet, blog, and journal club shared Cochrane evidence on a child health topic (e.g. treatments for asthma, prevention of childhood obesity). The social media strategy will be evaluated in the following ways: (1) Twitter and blog site analytics –measuring engagement with tweets, and blog site visits; (2) bit.ly statistics –measuring interaction with URL links; (3) Altmetrics –data on the change in scores of social media engagement with source evidence after our promotion; and (4) participant feedback on the journal clubs. We are also tracking the time and skill-sets required to develop the dissemination materials and manage the social media strategy. Results: A new blog post was uploaded on Monday of each week from November 2014 through March 2015 (22 weeks). We posted tweets three times per day; the tweets linked to the blog or to additional resources on the weekly blog topic. We ran journal clubs via Twitter each month from January to March. Results of the evaluation will be presented at the Colloquium. Conclusions: This work will provide empiric evidence for the utility of specific social media strategies for the dissemination of Cochrane evidence to professionals providing health services to children and youth. The results will provide useful information to others interested in developing a social media strategy or understanding the uptake of their social media activities. Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. O 6.3 Partnerships to co-ordinate and evaluate dissemination efforts Ruotsalainen J1 , Morata T2 , Lum M2 , Stewart G3 , Verbeek J1 1 Cochrane Work, Finland; 2 National Institute of Occupational Safety and Health, USA; 3 Wiley, United Kingdom Background: Dissemination of review results is one of the core objectives of Cochrane. With limited resources, how do we ensure maximum reach and impact for our reviews? Cochrane Work developed a new structured outreach strategy that is based on aligning dissemination efforts with domestic and international partner organizations. We aim to reduce repetition, expand reach and ultimately improve the uptake of review results. Objectives: We report on the initial phase of our co-ordinated dissemination efforts in partnership with the National Institute of Occupational Safety and Health (NIOSH) in the USA. Methods: We evaluate our concerted social media delivery with bibliometric and Altmetric data. To prepare for the release of the Cochrane Work review ‘Preventing occupational stress in healthcare workers’ we co-authored a NIOSH Science Blog and tweeted actively. We examined the effects of our dissemination efforts on blog views, number of tweets and review downloads. Results: The NIOSH Science Blog reached 1149 views from 960 unique visitors between 10 December 2014 (release date) till 24 March 2015. The Altmetrics score for this review was 36 and included 32 tweets from 28 accounts with a total reach of 28,480 combined followers. Plotting the tweets along the same timeline as the number of downloads did not reveal a clear pattern. While the number of downloads only slightly increased after publishing the blog, the Altmetrics score was among the top 5% of the scores in the Cochrane Library. Conclusions: While it is very early to evaluate our strategy beyond extended reach, partnering with institutions outside Cochrane improves dissemination efforts. It is feasible to use review downloads and Altmetrics scores as performance measures for review dissemination. O 6.4 RITES: a new systematic review tool to assess the characteristics of included trials along the efficacy/effectiveness continuum Wieland LS1 , Berman BM1 , Altman DG2 , Barth J3 , Bouter LM4 , D’Adamo C1 , Grimshaw J5 , Linde K6 , Moher D7 , Mullins D1 , Treweek S8 , van der Windt D9 , Merrick Z5 , Witt C6 1 Cochrane Complementary Medicine Field, USA; 2 Cochrane Bias Methods Group, United Kingdom; 3 Cochrane Complementary Medicine Field, Switzerland; 4 Cochrane Complementary Medicine Field, The Netherlands; 5 Cochrane Effective Practice and Organisation of Care Group, Canada; 6 Cochrane Complementary Medicine Field, Germany; 7 Cochrane Bias Methods Group, Canada; 8 Cochrane Methodology Review Group, United Kingdom; 9 Cochrane Complementary Medicine Field, United Kingdom Cochrane Database Syst Rev Suppl 1–327 (2015) 50 Background: Trials may address efficacy, i.e. intervention effects under ideal conditions, or effectiveness, i.e. intervention effects in real world circumstances. When systematic reviews are used to inform clinical or policy decisions, it is important to understand how applicable the evidence from the included trials is to the decision-maker’s context. Authors of systematic reviews can help by explaining where included trials are on the efficacy/effectiveness continuum. Objectives: To develop a valid, comprehensive tool that informs systematic review readers about the placement of individual trials on the efficacy/effectiveness continuum. Methods: We identified all relevant existing tools, then extracted the domains and the advantages and limitations of each tool as described in the literature. We conducted two online surveys with 72 stakeholders and organized two teleconferences and written consultations with 10 methodological experts to condense the domains and develop a draft of the tool. The feasibility and inter-rater reliability of the tool is being tested by piloting the tool with 10 trials included in three recent Cochrane complementary medicine reviews. The tool will be refined after piloting. Results: The RITES (Rating Included Trials along the Efficacy-effectiveness Spectrum) tool consists of four domains: (1) participant characteristics; (2) trial setting; (3) flexibility of interventions; and (4) clinical relevance of interventions. Evidence from the trial is rated in each domain on a 5-point Likert scale along a continuum from maximum emphasis on efficacy to maximum emphasis on effectiveness. Piloting is ongoing and results will be presented. Conclusions: RITES will help users of systematic reviews assess the ways in which the evidence from included trials provides pertinent information about effectiveness or efficacy. This will aid in the appropriate application of systematic review evidence in clinical practice or policy decisions. Rapid oral session 7 Quality of reporting RO 7.1 How often is a Cochrane review the best available evidence? Analysis of the more relevant clinical questions Rada G1 1 Epistemonikos Foundation, Chile Background: Relying on Cochrane Reviews (CRs) is recognized as an efficient way of making evidence-based health decisions. Objectives: To explore how often CRs are available for clinical questions; if they outperform other systematic reviews (SRs); and if they include all the relevant evidence. Methods: In the context of a project aiming to create friendly summaries of evidence, we established groups composed of at least one content expert in 15 departments of the faculty of Medicine of the Pontificia Universidad Católica de Chile. Each team prioritized at least five questions about Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. interventions that were considered hot topics in the area. We prepared matrices of evidence in Epistemonikos (a table comparing all of the reviews and all of the primary studies for a question). We measured the number of available SRs for each question; the number of questions answered by a CR/non-Cochrane review (NCR); and the number of times a CR was the newer/more complete. Results: Fifty-two clinical questions have been summarized at the moment. The average number of SRs per question was five (range 0 to 26). For 43/52 (83%) we identified at least one SR, but only 24/52 (46%) were answered by a CR. For the 24 questions for which a CR was available, in 16/24 (67%) the CR was the more complete, and in 12/24 (50%) it was the more recent. A total of 138/172 (80%) of the studies included in the CRs are included in other reviews. These results will be updated one month before the Colloquium. Conclusion: A substantial proportion of questions can be answered with SRs. However, non-Cochrane reviews (NCRs) had better coverage than CRs. When there were both CRs and NCRs, CRs were more complete, but there is ample room for improvement. A large proportion of the included studies in a CR can be obtained from existing reviews. Our results emphasize the need for producing reviews covering the more relevant topics and the challenges for Cochrane to be the more reliable source of evidence. Using existing SRs might be an efficient way to filter the information overload and to position CRs as the best available evidence. RO 7.2 Discrepancies in outcome reporting exist between protocols and published oral health Cochrane Reviews Pandis N1 , Fleming PS2 , Worthington H3 , Dwan K4 , Salanti G5 1 University of Ioannina & University of Bern, Switzerland; 2 Barts and The London School of Medicine and Dentistry, Queen Mary University of London, UK; 3 Cochrane Oral Health Group, School of Dentistry, The University of Manchester, UK; 4 Department of Biostatistics, University of Liverpool, UK; 5 Department of Hygiene and Epidemiology, University of Ioannina School of Medicine, Greece Background: In systematic reviews, selective reporting may develop for a variety of reasons, for example, due to the use of multiple measurement scales, outcomes or time points and selective inclusion of specific outcomes. There is paucity of information on selective reporting based on comparisons between review protocols and final systematic review reports. Objectives: To assess discrepancies between protocols and published reviews in oral health systematic reviews (COHG) on the Cochrane Database of Systematic Reviews (CDSR). Methods: Information on the reported outcomes from all COHG systematic reviews in CDSR and the corresponding protocols was recorded by two reviewers independently. Results: One-hundred and fifty-two reviews were included; the median number of labelled primary and secondary outcomes was two (range: 0 to 11) and four (range: 0 to 36) respectively both in the protocols and reviews. For primary outcomes, 11.2% (17/152) were downgraded Cochrane Database Syst Rev Suppl 1–327 (2015) 51 to secondary outcomes, 9.9% (15/152) were omitted and 18.4% (28/152) were introduced in the final publications. For secondary outcomes, 2.0% (3/152) were upgraded to primary, 12.5% (19/152) were omitted and 30.9% (47/152) were newly introduced in the publication. Overall, 45.4% (69/152) of reviews had at least one discrepancy that was justified in 14.5% (10/69) reviews. Sixty-three reviews included meta-analyses. For primary outcomes the risk of reporting significant results was lower for both downgraded (RR 0.52, 95% CI 0.17 to 1.58; P value 0.24) and upgraded or newly introduced outcomes (RR 0.77, 95% CI 0.36 to 1.64; P value 0.50) compared to outcomes with no discrepancies. For primary outcomes the risk for reporting significant results was higher for upgraded or newly introduced outcomes compared to downgraded outcomes (RR 1.19, 95% CI 0.65 to 2.16; P value 0.57). None of the comparisons reached statistical significance. Conclusion: There is evidence that discrepancies between outcomes of pre-published protocols and final reviews continue to be common, on the basis of this analysis of SRs published within the COHG. Alternative solutions to reduce the prevalence of this issue may need to be explored. RO 7.3 Shedding light on the maze of study labels and features: analysis of CBA and ITS studies in Cochrane Systematic Reviews Polus S1 , Pieper D2 , Fretheim A3 , Burns J1 , Ramsay C4 , Pfadenhauer L1 , Rehfuess E1 1 Institute for Medical Informatics, Biometry and Epidemiology, University of Munich, Germany; 2 Institute for Research in Operative Medicine, Department for Evidence-Based Health Services Research, Witten/Herdecke University, Germany; 3 Global Health Unit, Norwegian Knowledge Centre for the Health Services, Oslo, Norway; 4 Health Services Research Unit, University of Aberdeen, Foresterhill, Aberdeen, United Kingdom Background: Increasingly, Cochrane Systematic Reviews draw on nonrandomized study designs when assessing the effectiveness especially of complex interventions. In this context, controlled before-after (CBA) and interrupted time series (ITS) studies receive particular attention due to their recognition by the EPOC Group. There is, however, much confusion as a result of inconsistencies in study design labels and features and in relation to risk of bias. Objectives: To analyse CBA and ITS studies included in Cochrane Systematic Reviews descriptively in terms of their publication, application and methodological characteristics, focusing on strengths and weaknesses of the two designs. Methods: We searched the Cochrane Library for systematic reviews including nonrandomized studies from June 2012 to March 2015, updating a methodological study by Ijaz 2014. First, we classified the reviews including CBA or ITS studies according to ten pre-specified types of intervention. Subsequently, and separately for ITS and CBA studies, we purposively selected two reviews from each type of intervention. From each of Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. these reviews, we randomly sampled two CBA or two ITS studies, respectively. Data extraction and analysis were conducted separately by two authors using distinct tools for the two designs. Results: Our searches identified 126 reviews. From the sample of 40 CBA and 40 ITS studies, we described publication details, field of application (i.e. type and level of intervention), methodological characteristics (i.e. data collection, analysis), and quality assessment. The descriptive analysis yielded notable differences in relation to field of application of the two designs across health policy, health systems, behavioural, environmental, occupational, clinical, pharmaceutical, nutrition, screening and vaccination interventions and uncovered differences in quality of study conduct and reporting. Conclusions: This study provides further proof of heterogeneity in definitions and study features associated with the ’CBA’ and ’ITS’ labels. The strengths and weaknesses of the two designs have important consequences for reporting guidelines and quality appraisal tools. RO 7.4 Systematic review of adverse effects of antidepressants in healthy volunteer studies Bielefeldt AØ1 , Danborg PB1 , Gøtzsche PC1 1 The Nordic Cochrane Centre, Denmark Background: Clinical trials are often not published and adverse effects are often under-reported. Access to unredacted clinical trials is vital for independent researchers. In 2011 the Nordic Cochrane Centre got access to unredacted antidepressant trials from the European Medicines Agency (EMA), including 29 trials conducted in healthy volunteers. We wished to examine adverse effects (AEs) of SSRIs in healthy volunteers. Objectives: 1. A systematic review and meta-analysis of AEs of selective serotonin reuptake inhibitors in healthy volunteer studies, particularly activating effects and other AEs that can predispose to suicide. Specific for EMA trials: 2. A comparison of unredacted CSRs with published articles to investigate the selectivity of reporting. 3. A comparison of protocols with CSRs to investigate if protocol planned assessments were carried out. Methods: Searching for supplemental studies on PubMed and Embase for the systematic review. Eligibility criteria were set up. Studies were screened and relevant data extracted. Descriptive statistics were used to describe the included studies, and Peto OR for meta-analysis. Results: A systematic review of ∼150 healthy volunteer studies on antidepressants showed under-reporting of AEs. About one-third of studies did not mention AEs or their absence; 14 studies were eligible for meta-analysis, this showed that the incidence of AEs doubled when compared to placebo (Peto OR 2.20, 95% CI 1.35 to 3.59; P < 0.002). Only half the EMA trials (N = 29) were published (N = 15), and a mere few of these detailed AEs data fully (N = 2). Nine published articles had investigated AEs in their CSRs, but less than half (N = 4) reported these AEs, though these four studies all reported AEs partially. Result for objective 3 is pending. Conclusions: We investigated the AEs of antidepressants in Cochrane Database Syst Rev Suppl 1–327 (2015) 52 healthy volunteer studies. A systematic review of about 150 studies shows that about a third of all trials omit AE data and the meta-analysis shows that the incidence of the AEs possibly predisposing to suicide is doubled. Analysis of the EMA trials shows selective publication and inadequate reporting of AEs. RO 7.5 Agreement in results data between conference abstracts and full reports of randomized controlled trials: should we depend on conference abstracts? Saldanha I1 , Scherer R1 , Dickersin K1 1 Cochrane Eyes and Vision Group, USA Background: Including conference abstracts in systematic reviews promotes comprehensiveness, but conference abstracts are not usually peer-reviewed and often contain preliminary data. Conclusions about intervention efficacy in randomized controlled trials (RCTs) are generally based on primary outcomes (POs). Objective: To quantify agreement between PO results of RCTs presented as conference abstracts and their corresponding full reports. Methods: We included all abstracts of RCTs presented at the 2001-2004 Association for Research in Vision and Ophthalmology conferences. We identified corresponding full reports through to 2013 by Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. electronic searching and emailing authors. We extracted data about the PO from each abstract and full report. For each abstract-full report pair where the PO was the same and direction of results (positive or neutral) could be determined in both publications, we examined whether the results agreed. We classified any discordance as quantitative (any change in magnitude but not direction of effect), qualitative (change in direction of effect), or both. Results: Two-hundred and thirty (44.8%) of the 513 eligible abstracts had been published in full. Within these 230 abstracts, direction of results for the PO was positive for 64 (27.8%), neutral for 55 (23.9%), and could not be determined for 119 (51.7%). POs differed in 82/230 pairs (36.9%). POs were more likely to differ if the abstract’s results were neutral compared to positive (RR = 1.28, 95% CI 1.15 to 1.40). Among the 103 abstract-full report pairs in which the PO was the same and direction of results could be determined, results agreed in 20 (19.4%), there was quantitative discordance in 74 (71.9%), and both quantitative and qualitative discordance in 9 (8.7%; Figure 1). Conclusions: POs in more than one-third of RCTs presented as abstracts differed from POs in the full reports; a difference was more likely if the direction of PO results in the abstract was neutral. When the POs were the same in both, only one-fifth of pairs agreed, and almost one-tenth of pairs had qualitatively different results. Systematic reviewers should be aware of, and cautious about, these differences. Cochrane Database Syst Rev Suppl 1–327 (2015) 53 RO 7.6 Methods for data extraction from figures in protocols of Cochrane Systematic Reviews Vucic K1 , Jelicic Kadic A2 , Puljak L2 1 Agency for Medicinal Products and Medical Devices, Croatia; 2 Cochrane Croatia, Croatia Background: When conducting a systematic review, some data that need to be extracted from primary studies are presented only as figures. One can contact study authors to request original data from figures, but those requests for data seldom yield a response. In those circumstances data need to be extracted from figures. However, the Cochrane Handbook does not cover issues relating to data extraction from figures. Objectives: To analyze whether protocols of Cochrane Systematic Reviews mention whether and how the authors will extract data from figures in included trials. Methods: Cochrane Protocols published between May 2013 and May 2014 were screened by two authors independently and the following data were collected: date of protocol publication, country of authors’ origin, number of authors, number of affiliated institutions, Cochrane Review Group, whether the protocol contains description of data extraction from figures, the method of data extraction from figures, and literature references for a method of data extraction from figures. Results: From a total of 589 protocols, 33 (5.6%) mentioned data extraction from figures in the Methods section. Most of those studies mentioning data extraction from figures (18/33) indicated that data from figures would be included only if two reviewers independently extracted the same result. Only one protocol specified that computer software would be used for data extraction from figures, and few stated estimation or approximation, while others did not provide any description of methodology for data extraction from figures. One protocol specifically indicated that data from figures would not be used, ‘‘because of [the] possibility of making measurement errors when estimating from graphs’’. Conclusion: Very few Cochrane Protocols mention data extraction from figures and, even when they do mention data that are presented only in figures, it is not necessarily to explain how those data will be extracted. Some even plan to disregard those data completely. The Cochrane Handbook and methodological standards for Cochrane Systematic Reviews should address methodology for extracting data from figures. RO 7.7 Reliability of data extraction using manual measurements of published graphs or figures Pérez-Gaxiola G1 , Cuello-Garcia C2 1 Cochrane México, Hospital Pediátrico de Sinaloa, Mexico; 2 Department of Clinical Epidemiology and Biostatistics, McMaster University, Canada Background: Meta-analyses combine results of specific outcomes from individual primary studies, in which trial Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. authors occasionally publish results with figures or graphs without the exact numerical data needed for further analysis. Although the most appropriate way to deal with missing numbers is to contact authors, sometimes this is not feasible. One alternative is to calculate manually by extracting numbers using measurements in figures as a proxy. Objectives: To analyse the reliability/accuracy of data extraction using manual measurements from figures published in individual studies. Methods: Four figures illustrating a clinical outcome without an exact numerical value embedded in them were selected. The corresponding published study did contain an exact value that we used as gold standard. We asked participants to extract from the figures the value of the clinical endpoint of interest using ruler measurements within a PowerPoint slide as a proxy. Raters were blinded to the gold standard. We measured inter-rater reliability of participants using Type A intra-class correlation (ICC) coefficients with absolute agreement definition; we also compared the mean from the raters for each figure to their correspondent gold standard. Results: Twenty-nine raters agreed to participate. In the four figures, less than 1% of the measurement corresponded to random variation (ICC coefficient 0.99, 95% CI 0.96 to 1.0). Only in one figure there was a statistically significant difference between the participant-estimated mean and the gold standard. Conclusions: Measurement of an estimate of interest that is not displayed on figures from individual studies could be extracted from them by hand using any presentation program that a methodologist is familiar with. As with any extraction process in a review, extraction by duplicate is warranted to ensure exactitude of the final estimate. RO 7.8 A graphical display of reporting factors associated with systematic reviews: an example from diabetic retinopathy Hui X1 , Clearfield E1 , Lindsley K1 , Virgili G2 , Scherer R1 1 Cochrane Eyes and Vision Group, Johns Hopkins University, USA; 2 Cochrane Eyes and Vision Group, University of Florence, Italy Background: Systematic reviews (SRs) are considered to be the highest level of evidence, but may lead to biased conclusions if not properly conducted. We aim to evaluate SRs of interventions for diabetic retinopathy (DR), the leading cause of blindness in the working-age population in developed countries. Objectives: To assess the descriptive, methodological, quantitative, qualitative, and other characteristics of SRs of interventions for DR. Methods: We searched PubMed/MEDLINE, EMBASE and the Cochrane Central Register for Controlled Trials (CENTRAL) up to May 2014 for all SRs related to eyes and vision with a focused question and explicit scientific approach. We identified all SRs that examined an intervention for DR. Two authors independently evaluated 15 characteristics adapted from the Critical Appraisal Skills Programme (CASP) and A Measurement Tool to Assess Systemic Reviews (AMSTAR; Cochrane Database Syst Rev Suppl 1–327 (2015) 54 Figure 1 Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Cochrane Database Syst Rev Suppl 1–327 (2015) 55 Table 1). We noted whether each item was reported or not reported/can’t tell and calculated the proportion of SRs reporting that characteristic. We plotted the calculated proportions using a radar plot in MS Excel (Figure 1). Results: We identified 25 SRs published from 2002 to 2014; four were Cochrane SRs. The SRs included a median of six studies (range: 0-11). Common interventions included anti-vascular endothelial growth factor (VEGF) treatments (14 trials) and corticosteroids (10 trials). The majority of SRs presented eligibility criteria (96%), however only 36% reported that the criteria were pre-specified. Only 56% of searches were comprehensive. A risk of bias assessment was performed in 52% of SRs, and 61% discussed limitations. Seventy-nine per cent used correct methods when results were combined quantitatively. Non-industry or no funding was reported by 56%. Conclusions: The radar graph visually displays substantial gaps in the reporting of descriptive factors and appropriate SR methodology for SRs in DR. Reporting of quantitative factors was slightly better compared with qualitative methods, such as discussing limitations. types. Outcomes related to clinical or nutritional status assessment were most frequent (n = 385; 96%), followed by disease incidence or prevalence (n = 199; 49%). Nonrandomized studies were included in only three reviews (< 1%) and GRADE was utilized in 85 (21%). The quality assessment is in progress and will be reported at the Colloquium. Conclusions: A relatively small proportion of Cochrane Reviews address nutrition-related interventions. These are mainly concentrated in two CRG’s, with remaining reviews thinly spread across other groups. Current reviews focus on nutritional supplements and clinical or nutritional status outcomes. These findings can inform Cochrane actions to improve the quantity, quality and relevance of nutrition-related reviews for better nutrition policies and practice. Rapid oral session 8 Evidence maps and priority setting RO 8.1 RO 7.9 Scope and quality of Cochrane Reviews of nutrition-related interventions Harper A1 , Naude CE2 , Durao S3 , Volmink J4 1 Division of Human Nutrition, Stellenbosch University, South Africa; 2 Centre for Evidence-based Health Care, Stellenbosch University, South Africa; 3 South African Cochrane Centre, South African Medical Research Council, South Africa; 4 South African Cochrane Centre, South African Medical Research Council; Centre for Evidence-based Health Care, Stellenbosch University, South Africa Background: Despite the importance of nutrition for health and development, a limited number of nutrition-related Cochrane Reviews exist. These reviews are of uncertain quality and relevance. Objectives: To assess the scope and methodological quality of nutrition-related reviews in the Cochrane Database of Systematic Reviews (CDSR). Methods: We screened the CDSR (February 2014) to identify nutrition-related reviews and protocols according to pre-specified eligibility criteria. Two authors independently performed data extraction (PICOS) and quality assessment (AMSTAR) of eligible reviews. We report our findings using statistical and narrative description. Results: We screened 8217 records, which yielded 536 (6.5%) relevant reviews (n = 400) or protocols (n = 136) distributed across 45 Cochrane Review Groups (CRGs). Eight CRGs had no nutrition-related reviews. The median number of nutrition-related reviews across all CRGs was four (interquartile range: 1 to 9). CRGs with the most reviews were Pregnancy & Childbirth (n = 64/527), Neonatal (n = 60/303), Metabolic & Endocrine Disorders (n = 26/100), and Developmental, Psychosocial & Learning Disorders (n = 23/121). The most common type of intervention evaluated was nutritional supplements (n = 232), followed by food/dietary patterns (n = 139), with some reviews including several intervention Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. EBRNetwork –a call to action for more (efficient) systematic reviews Nasser M1 , Westmore M1 , Lund H2 , Brunnhuber K1 , Dobbins M3 , Robinson K4 , Leenars M5 , Nykvist H6 , Christensen R2 , MacLeod M1 , Dorch B2 , Ford P1 1 Evidence-Based Research Network, United Kingdom; 2 Evidence-Based Research Network, Denmark; 3 Evidence-Based Research Network, Canada; 4 Evidence-Based Research Network, USA; 5 Evidence-Based Research Network, The Netherlands; 6 Evidence-Based Research Network, Norway Background: Initiating new research projects without reviewing the existing evidence systematically is wasteful, unscientific, and potentially unethical. Scientists are expected to refer to earlier research results when they argue for the need for a new study. However, research-on-research shows that the papers cited are often insufficient and biased towards the interest of the researchers; that researchers are not supported (through funding, time, training) in the production and updating of systematic reviews; and that there is a need for new ways of conducting mandatory systematic reviews that are more efficient yet rigorous. Objectives: To address these problems, a group of researchers have initiated an international network, the ‘Evidence-Based Research Network’ (EBRNetwork - www.ebrnetwork.org), established in Bergen, Norway in December 2014. Methods: We are presenting the findings of several relevant studies, including the use of previous research by scientists; the problem of ongoing research after a benefit or harm of an intervention has been unequivocally established; and an international comparison of research funders, regulators and publishers regarding policies mandating systematic reviews prior to new research. We issue an invitation to join the EBRNetwork and will work with the members of the audience toward identifying and prioritising key initial workstreams. Results: Researchers embarking on research after reviewing Cochrane Database Syst Rev Suppl 1–327 (2015) 56 the existing evidence systematically will be able to clarify whether the study is truly needed (adding value to health care),or whether it is asking a research question for which we have already have high quality evidence –and thus could be considered wasteful. Conclusions: The new EBRNetwork is an international collaboration that aims to ensure that no new studies are conducted without prior systematic review, and that works towards more efficient production, updating and dissemination of systematic reviews. The Network issues a call to participate in developing a consensus statement to accomplish these aims. RO 8.2 Evidence gap maps –a tool for promoting evidence informed policy and prioritising future research Snilstveit B1 , Vojtkova M1 , Stevenson J1 , Bhavsar A1 , Gaarder M2 , Phillips D1 , Gallagher E1 1 International Initiative for Impact Evaluation, United Kingdom; 2 World Bank, United Kingdom Background: A range of organisations are engaged in the production of evidence on the effects of social and economic development interventions. The growth in the production of evidence also presents challenges. How can decision makers avoid information overload and access the best available evidence in a specific field? How can we present research in a format that is useful and accessible for a non-technical audience? And how can we ensure limited resources are spent efficiently and important evidence gaps are prioritised? Objectives: Evidence gap maps (EGMs) aim to respond to these challenges and present a new addition to the tools available to support evidence-informed policy making. The objective of this project was to develop a method for evidence mapping and visualisation to aid knowledge translation and strategic research prioritisation. Methods: We reviewed the methodological literature on evidence mapping and synthesis to identify existing methods to identify, appraise and present evidence from primary studies and systematic reviews. We designed a visual framework for presenting the evidence and worked with web developers to create an interactive and visual platform. Results/Conclusions: EGMs provide a visual overview of what we know and don’t know about the effects of interventions in a sector or sub-sector. EGMs are based on systematic methods and are structured around a framework which schematically represents the types of interventions and outcomes of relevance to the thematic area. They enable policy-makers and practitioners to explore the findings and quality of the existing evidence to facilitate evidence informed decision-making. EGM also identify important ’gaps’ where little or no evidence from impact evaluations and systematic reviews is available, and can inform a strategic approach to Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. building the evidence base in a particular sector. The paper provides an introduction to the concept of evidence gap maps and their methodology, and a demonstration of the EGM tool using existing examples. RO 8.3 Evidence gap maps for eye health: tool for promoting evidence and gaps in research Virendrakumar B1 , Jolley E1 , Gordon I2 , Bascaran C3 , Pellegrino C1 , Schmidt E1 1 Sightsavers, United Kingdom; 2 Cochrane Eyes and Vision Group, London School of Hygiene & Tropical Medicine, United Kingdom; 3 London School of Hygiene & Tropical Medicine, United Kingdom Introduction: Sightsavers is an international non-governmental organisation (NGO) working to eliminate avoidable blindness in low- and middle-income settings (LMIS). High quality evidence is essential for planning and delivering our programmes, however, evidence on what works to improve eye health in LMIS is relatively scarce and there is no single repository where all relevant research may be found. To address this need, Sightsavers is developing eye health Evidence Gap Maps (EGMs) using the methodology developed by the International Initiative for Impact Evaluation (3ie). EGMs summarize, critically appraise and present evidence from systematic and literature reviews on a particular topic in a user-friendly format. The quality appraisal is integrated into the visual presentation of the EGM and shows where the strengths and weaknesses of evidence lie. We will present our experiences of developing the cataract EGM including the challenges and benefits along with the initial evaluation of its potential to inform the development of EGMs for other eye diseases. Methods: Following a comprehensive search of relevant databases, websites and reviews of references, we sifted, and extracted data from all relevant reviews on cataract. Critical appraisal was conducted by two reviewers independently using Supported Use of Research Evidence (SURE) checklist and a summary of quality assessment was shared with the authors for comment. Results: Out of 1197 unique studies identified, 52 reviews met the inclusion criteria. Most reviews (18) addressed surgical methods for cataract, followed by quality of clinical care (17). Of the remainder, seven looked at risk factors/prevention, five studies were impact/economic evaluations, and three each addressed the burden of the disease and the accessibility of cataract services. Twenty-five studies were graded as being of low methodological quality, and 13 as being high quality. Discussion: EGMs will enable decision makers to make more informed choices by presenting the effectiveness of eye health interventions in a simple visual format. They will also enable researchers to prioritise areas where little evidence exists. Cochrane Database Syst Rev Suppl 1–327 (2015) 57 RO 8.4 Accountability and health information systems: from health systems’ priority setting to a pragmatic systematic review Leon N1 , Brady L1 , Kwamie A2 , Lewin S3 , Odendaal W1 , Daniels K1 1 South African Medical Research Council, South Africa; 2 School of Public Health, University of Ghana, Ghana; 3 Norwegian Knowledge Translation Centre for Health Services and Health Ssytems Research, Norway Background: The Alliance for Health Policy and Systems Research recently funded a two-year initiative in South Africa to conduct four systematic reviews on health systems questions. Central to the initiative was engagement with stakeholders to identify and select priority review questions. While the priority setting process was rich and very broad, these broad questions were not always easily translated into questions that could be addressed by a review with limited budget and timeframes. Objective: To share lessons learned in turning a broad health systems policy priority question about effective accountability interventions, into a pragmatic systematic review question. Results: Through the process of stakeholder engagement 39 priority review questions were shortlisted. The four highest-ranking questions for which no published or ongoing review existed were selected for further investigation. One of these questions focused on the effectiveness of interventions to improve the accountability of public sector health personnel and management. While enhanced accountability is important to policy makers and managers, we soon realised the broad and abstract nature of this concept raised challenges in relation to definition, operationalisation and measurement. Through a process of literature review, consultation and discussion we explored the components of accountability, and how it might be operationalised within the health system. This process led us to shift our focus to health systems performance monitoring, as a proxy measure of accountability. The result was a revised review question on the effectiveness of interventions to improve health information systems. Improved health information systems are an essential tool for measuring progress and accountability and for filtering information for better decision-making. Conclusion: Policy priorities as articulated by stakeholders may not be easily translated into systematic review questions. Thus flexibility and ongoing dialogue are required between researchers and policy stakeholders when identifying health systems systematic review topics collaboratively. RO 8.5 Evidence gaps in health policy and systems research in 15 countries of the Eastern Mediterranean Region: a mapping exercise El-Jardali F1 , Bou-Karroum L1 , Akl E1 1 American University of Beirut, Lebanon Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Objective: This study aims to map the gaps in the production of Health Policy and Systems Research (HPSR) evidence in 15 countries of the Eastern Mediterranean Region (EMR). The study also assesses the alignment between the existing HPSR and high-level regional and global priorities. Methods: We searched MEDLINE for articles on HPSR published between 2000 and 2013 in the 15 EMR countries (Bahrain, Egypt, Jordan, Kingdom of Saudi Arabia, Kuwait, Lebanon, Libya, Morocco, Oman, Palestine, Qatar, Sudan, Syria, Tunisia, and Yemen). Teams of two reviewers assessed eligible articles using a coding form that included HPSR themes on governance, financial and delivery arrangements and implementation strategies. We then matched articles identified as HPSR to regional and global priorities pertaining to 11 themes including health human resources, health financing, role of non-state sector, access-to-medicine, primary healthcare, non-communicable diseases, universal health coverage, emergency preparedness and response. Results: Out of the 29,126 articles published in the 15 EMR countries, 9% fitted the criteria for HPSR. There was an increase in production after 2005. HPSR articles focused on themes of delivery arrangements (68%) and implementation strategies (19%). The evidence gap was noted in the themes of financial (3%) and governance arrangements (9%). We found misalignment between HPSR produced in the region and the regional and global priorities. Conclusions: The mapping of evidence gaps persistently showed low production of HPSR in the region. The mapping should inform the research agenda (for researchers and funders) for the field, including identifying review questions for systematic reviews. It also demonstrates the need for aligning the production of HPSR with policy needs and priorities, and for building the capacity in conducting HPSR at the individual, team, institutional and system level. RO 8.6 Establishing and prioritising a local health research agenda with end users Moore D1 , Abbott R1 , Rogers M1 , Bethel A1 , Stein K1 , Thompson-Coon J1 1 University of Exeter, United Kingdom Background: The National Institute for Health Research Collaboration for Leadership in Applied Health Research and Care in the South West of England (PenCLAHRC) is developing a portfolio of clinically relevant, locally tractable and patient-informed research projects through a process involving a wide group of end users at all stages, from the inception of research ideas through to delivery of outputs. Objectives: This paper describes and evaluates the research prioritisation process being used by PenCLAHRC; a partnership of local National Health Service (NHS) and public health organisations across Somerset, Devon and Cornwall, with the Universities of Exeter and Plymouth. Methods: PenCLAHRC identifies research questions from its partners, including members of the public. Questions are identified during engagement with service users, clinical Cochrane Database Syst Rev Suppl 1–327 (2015) 58 teams and organisations around the use of evidence and via the PenCLAHRC website. Questions received are prioritised in a process involving all PenCLAHRC stakeholders (including our Peninsula Patient and Public Involvement group). Prioritisation is based on a set of explicit criteria including importance, local relevance and feasibility. In 2014 we piloted a novel approach consisting of two rounds of electronic voting, before a face-to-face meeting to discuss and rank the prioritised questions. Results: Seventy-two initial questions were reduced to 50 and then to nine in two electronic rounds of comments and voting. Priority briefings for the nine questions were prepared and were discussed and voted on at a face-to-face meeting in December 2014. The process has been evaluated by stakeholders. Feedback indicates that stakeholders found voting electronically, sharing comments on questions and more than one round of voting valuable. Issues regarding the time available for prioritisation activities and the quality of some identified questions were raised. Conclusions: Engagement of end users in establishing a research agenda increases its relevance and may promote use of research. With minor changes, future rounds of question prioritisation will build upon the process outlined here. RO 8.7 A priority setting tool for prioritizing review questions for systematic reviews in health policy and systems research: development and validation Akl E1 , El-Jardali F1 , Fadlallah R1 1 American University of Beirut, Lebanon Background: Groups conducting systematic reviews in health policy and systems research need to prioritize topics according to the needs of policymakers and stakeholders. Such prioritization can help ensure systematic review production is aligned with policy priorities, which, in turn, can promote their use in policy making and practice. A systematic search did not identify a tool to support such prioritization process. Objectives: The aim of our study is to develop and validate a tool for prioritizing questions for systematic reviews in health policy and systems research. Methods: The development and validation methodology includes the following three stages: 1. definition of the purpose and scope of the too; 2. tool development/item generation; and 3. validation of tool. The latter stage involves the following steps: 1. establishing content and face validity; 2. pilot testing of tool in real life priority setting exercises with policymakers and stakeholders from different contexts; 3. exploratory factor analysis; 4. reliability testing; 5. external construct validity; 6. refinement and finalization of tool; and 7. translation of tool into different languages. Results: We have completed the first two stages, and are currently working on the second step of the final stage. The tool includes two modules: one (14 items) to be completed by policymakers and stakeholders; and the second Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. (nine items) to be completed by systematic reviewers. In the presentation, we will provide a detailed description of the ’finalized’ tool, the process and outcome of its development, and how it could be used by systematic review teams and centers. Discussion will also delve into lessons learned and implications for other groups that conduct priority setting exercises for systematic reviews. Conclusions: The tool might be helpful for organizations in both prioritizing their systematic review work, and ensuring that they are used to inform policy decisions. RO 8.8 Priority setting for research and practice of child health in low- and middle-income countries (LMICs) based on evidence from Cochrane Reviews Sinha A1 , Ovelman C2 , Pradhan A1 1 Cochrane Neonatal Group, India; 2 Cochrane Neonatal Group, Canada Background: Systematic Reviews (SRs) identify, appraise and synthesize research-based evidence and present it in an accessible format for clinicians and healthcare practitioners. The knowledge emerging from SRs may be of help to health planners when setting priorities for health research, implementing proven interventions, and using limited resources judiciously. Objectives: To identify interventions with high-level evidence and research leads/primary research questions in child health emanating from Cochrane SRs relevant to low- and middle-income countries (LMICs). Methods: We searched six Cochrane Review Groups (Acute Respiratory Infections (ARI); Infectious Diseases; Neonatal; Cystic Fibrosis and Genetic Disorders; Airways; Developmental, Psychosocial and Learning Problems (DPLP)) that produce reviews relevant to child health in LMICs for published and updated Cochrane SRs from 1 March 2009 to 18 March 2015 in the Cochrane Library. We extracted data from the objectives, main results and authors’ conclusions sections of the Cochrane SRs to identify primary research questions. Results: We found a total of 679 Cochrane SRs (105 from ARI, 178 from Airways, 100 from Cystic Fibrosis, 78 from DPLP, 63 from Infectious Diseases, and 155 from Neonatal) addressing research questions of importance to LMICs. About 50% of these Cochrane SRs concluded that there was insufficient evidence for making valid conclusions (not enough RCTs, small sample sizes in included studies). Data on interventions with high-level evidence applicable to public health system in LMIC, and also data on lead research questions (from reviews with insufficient evidence) will be presented in the paper. Conclusions: Cochrane Reviews with children as subjects may be used to implement proven interventions, identify knowledge gaps to guide future research as per national child health priorities in LMICs facing resource constraints. Cochrane Database Syst Rev Suppl 1–327 (2015) 59 Rapid oral session 9 Meta-analysis methods RO 9.1 The selection of fixed- or random-effect models in recent published meta-analyses Kuan Y1 , Tam K1 1 Shaung-Ho Hospital, Taipei Medical University, Taiwan Background: Most meta-analyses are based on one of two statistical models. A fixed-effect meta-analysis assumes all studies are estimating the same (fixed) treatment effect, whereas a random-effects meta-analysis allows for differences in the treatment effect from study to study. The selection of a model must be based on the question of which model fits the distribution of effect sizes, and takes account of the relevant sources of error. Objectives: The study aims to evaluate the preference and selection of statistical model in recent published meta-analyses. Methods: The published meta-analyses were extracted from PubMed searches before 18 March 2015. We retrieved 60 studies and investigated their selection of statistical models. Results: Six of these 60 studies did not report whether fixed- or random-effects was used for meta-analysis. Random-effects pooling model were conducted in 27 meta-analyses. Both fixed- and random-effect models were used simultaneously in five studies. In another 22 studies, a fixed- or random-effect model was chosen according to the heterogeneity. For example, studies with an I2 statistic of > 50% were considered to have substantial heterogeneity, and therefore, a random-effects model analysis was used. Otherwise, a fixed-effect model was initially employed in the analysis. Interestingly, 21 of the 60 meta-analyses were reported from China, and 15 of them selected fixed- or random-effect models according to the heterogeneity. Conclusions: The fixed-effect model starts with the assumption that the true effect size is the same in all studies. However, in many systematic reviews this assumption is implausible. Therefore, when studies are gathered from the published literature, the random-effects model is generally a more plausible match. Unfortunately, 36.7 % of our included studies starting their analysis with a fixed-effect model and then moved to a random-effects model if the test for heterogeneity was significant. The above strategy is thought to be a mistake, and should be discouraged. RO 9.2 Empirical comparison of two methods for assessing interaction in aggregate-data meta-analysis of randomized controlled trials: a literature-based survey and analysis Yang Z1 , Mao C1 , Tang J1 1 The Hong Kong Branch of the Chinese Cochrane Centre, China Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Many treatments are more effective in some patients than others, which means that the effectiveness of treatment varies with some factors, usually called ’interaction’. There are several methods available for investigating interaction in aggregate-data meta-analyses. Suppose we want to examine the potential interaction between sex and treatment, using relative risk (RR) as the outcome measure. The best method (’interaction term approach’) would be to calculate an interaction term based on RR-male and RR-female within each trial, and then combine the interaction terms across trials to obtain a summary interaction term. However, the majority of existing meta-analyses do not use this method. Instead, they combined the treatment effects in male subgroups across trials first (RR-maletotal), and then combined the treatment effects in female subgroups (RR-femaletotal), and then test for difference between the two subgroup-specific summary estimates (’subgroup approach’). This method can be easily implemented with commonly available software for meta-analysis such as RevMan. However, the subgroup approach is problematic. First, it calculates the subgroup-specific summary estimates first, thus ignoring the within-trial correlations between male and female in the trials with data on both subgroups. This would lead to reduced efficiency and inappropriate estimation of standard errors. In addition, to compare the two summary estimates of male and female subgroups, the heterogeneity in the meta-analyses to obtain the two summary estimates has to be ignored. For these reasons and based on the conceptual comparisons, researchers have recommended avoiding using the subgroup approach. In this study, we searched PubMed and randomly selected 100 empirical meta-analyses to re-evaluate the interactions, which were originally assessed by the subgroup approach, by using the interaction term approach. The results on interactions by the two methods were compared both qualitatively and quantitatively. Study characteristics associated with differences in the results were described. Recommendations were made for future meta-analyses. Details will be presented at the meeting. RO 9.3 An investigation of the type I error rate when testing for subgroup differences in the context of random-effects meta-analyses Guddat C1 1 IQWiG, Germany Background: There are different approaches to test for differences between two or more subgroups of studies in the context of a meta-analysis. The Cochrane Handbook for Systematic Reviews of Interventions refers to two methods. One is a standard test for heterogeneity across subgroup results rather than across individual study results. The second is to use meta-regression analyses. Objectives: Our aim was to compare the performance of these two approaches with respect to the type I error rate when 10 or fewer studies are available. Methods: Assuming the Cochrane Database Syst Rev Suppl 1–327 (2015) 60 random-effects model, we have conducted a simulation study for the planed comparison. Two versions of the test for heterogeneity have been considered: one with separate estimates of the between-study variance in each subgroup, and one with a pooled estimate of this variance among all subgroups. The meta-regression was conducted using two modifications for the variance estimator of the regression coefficients by Knapp and Hartung (2003). Besides the number of studies, we varied, amongst other parameters, the extent of heterogeneity between the studies, the number of subgroups and the distribution of the studies to the subgroups. Results: Both versions of the test for heterogeneity give extremely high error rates when the heterogeneity between the studies is large and the distribution of the studies to the subgroups is unequal. In contrast, the error rates of the F-test using either of the two considered variance estimators in the context of a meta-regression are acceptable, irrespective of the chosen parameters. Conclusions: Overall, the F-test using the refined variance estimator by Knapp and Hartung (2003; 1. modification) is the most appropriate choice out of the evaluated tests with respect to the type I error rate when the number of studies is 10 or fewer. References: Knapp G, Hartung J. Improved tests for a random effects meta-regression with a single covariate. Statistics in Medicine 2003: 22(17): 2693–2710. RO 9.4 Meta-analyzing correlation coefficients derived from cohort studies Thumburu KK1 , Singh M2 , Kaur J1 , Singh S1 , Jaiswal N1 , Chauhan A1 , Agarwal A1 , Paul N3 , Sagwal S3 1 ICMR Advanced Centre for Evidenced-Based Child Health, Advanced Pediatric Centre, PGIMER,Chandigarh, India; 2 Department of Pediatrics and ICMR Advanced Centre for Evidenced-Based Child Health, Advanced Pediatric Centre, PGIMER,Chandigarh, India; 3 Department of Pediatrics, Advanced Pediatric Centre, PGIMER,Chandigarh, India Background: Correlation coefficient (r) is a measure of strength and direction of linear association between two variables or used as quantitative surrogate of individual variables. There have been no or few studies in which pooling of r values has been done and hence, we are presenting a meta-analysis of non interventional cohort studies by combining the r values. Objectives: To meta-analyze correlation coefficients derived from cohort studies. Methods: We conducted a systematic review on association of anthropometric measures during childhood and risk of becoming overweight or obese in later life. The studies included were prospective cohort or longitudinal studies. Tracking estimates such as the correlation coefficient are preferred compared to difference estimates when assessing body-mass index (BMI) tracking. Basic information recorded consisted of cohort size n, (mean) age at baseline, and at follow-up measurement. The effect size calculated was correlation coefficient (either Pearson’s or Spearman’s). Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Standard error was calculated using the information from cohort size and r values. Meta-analysis was done using inverse variance random-effects model. Stata/MP 12.2 software) was used to perform the meta-analysis. Results: Twenty-eight studies provided correlation coefficient data on BMI tracking from 40,219 individuals for meta-analysis. Follow-up time ranged from two to 65 years. We derived standard errors from the available sample size and r value. Then r values were pooled using the inverse variance random-effects model. The estimate size with confidence intervals provided the net effect. Conclusions: Pooling of correlation coefficients is a useful and feasible means of meta-analyzing quantitative data from cohort studies. Acknowledgment: The study was supported by World Health Organization, Geneva, Switzerland and ICMR, New Delhi, India RO 9.5 Joint models to account for informative losses to follow-up in longitudinal studies Brignardello-Petersen R1 , Singer L1 , Herridge M2 , Tomlinson G1 1 Institute of Health Policy, Management and Evaluation, University of Toronto, Canada; 2 Toronto General Research Institute, University Health Network, Canada Background: A major challenge in the analysis of longitudinal studies is data that are missing because participants’ follow-up ends before the planned follow-up period ends. Many times, these data are missing not at random, and use of conventional longitudinal models (e.g. linear mixed-effects models) can lead to biased results for the coefficients relating participants’ characteristics to outcomes. Joint models (JM) that simultaneously estimate time-to-event and longitudinal data models are a new alternative that account for informative losses to follow-up in longitudinal studies. Objectives: To explore whether the results of longitudinal studies with a high proportion of informative losses to follow-up change when the results are analyzed using JMs compared to linear mixed-effects models. Methods: We reanalyzed the data from two longitudinal studies with a high proportion of informative losses to follow-up due to patient death, using JM. In the first study the aim was to determine the trajectories of quality of life of patients after receiving a lung transplant. In the second study the aim was to assess the trends in function of patients discharged alive from the ICU, and determine whether these differed among risk groups. There was approximately 35% and 25% of informative losses to follow-up, respectively. The data was analyzed using the package JM in R. Results: The change in the primary outcome over time was statistically significant when analyzing the datasets using linear mixed-effects models. In the second example, the change in the primary outcome was also different among groups. Although in both datasets, the change in the slope over time was different when using JM to account for the informative losses to follow-up, there were no important changes in the statistical significance of the results for the main variables of interest. However, in both cases, the differences in the point estimates of the slopes were too small Cochrane Database Syst Rev Suppl 1–327 (2015) 61 to change the clinical conclusions drawn from the analysis. Conclusions: In these examples, the clinical conclusions from the longitudinal studies did not change when accounting for the informative losses to follow-up using JM. RO 9.6 The influence of mortality time-points on pooled effect estimates in critical care meta-analyses Roth D1 , Herkner H1 1 Medical University of Vienna, Department of Emergency Medicine, Austria Background: There is an on-going debate among meta-analysis methodologists and statisticians whether it is appropriate to pool mortality estimates from clinical trials that use mortality outcomes ascertained at different time-points. If the relative effects vary over time, which might especially be the case in critical care, standard pooling of studies with different follow-up times within one meta-analysis would not be justifiable. Current Cochrane guidelines propose pooling of short-, middle-, and long-term effects as a potential solution, but include no specific guidelines. Objectives: Describe the current practice of dealing with different mortality time-points and analyze the influence of different time points on pooled effect estimates in actual Cochrane critical care meta-analyses. Methods: The CDSR was searched for critical care-reviews. Review characteristics including strategy for dealing with different follow-up times and study characteristics were extracted. Meta-analyses were recalculated using all described strategies and influence of such strategies on deviation of pooled effect estimates compared to a ”use last time-point available” approach was analyzed using meta-regression and multilevel mixed-effects linear regression. Results: We evaluated 835 reviews, and included 80 meta-analyses of 298 studies, representing 107,605 patients. 49 (61%) reviews did not state any strategy, 9 (11%) used separate analyses for each time-point, 9 (11%) used the last available, 6 (8%) used a closest to defined time-point, 3 (4%) performed separate analyses for last and predefined, 2 (3%) mixed some, 1 (1%) computed predefined time-points from study-data, and 1 (1%) pooled all but performed a sensitivity analysis. Among 388 recalculated meta-analyses no influence of the strategies ”pool short-, middle-, long-term”, ”use closest to defined” and ”separate” on effect estimates was found compared to ”use last available”. Conclusions: Reviews use a large variety of strategies to deal with different mortality time-points, however more than 50% do not report any strategy. We found no influence of different strategies on effect estimates in critical care reviews. Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. RO 9.7 Multiple-arm trial data: using a corrected standard error for GIV analyses Cates C1 1 Airways Group, United Kingdom Background: Sometimes multiple-arm trials only report the pairwise difference between each arm and the control (with a 95% confidence interval (CI)). These paired differences can be individually entered into Review Manager 5 using the Generic Inverse Variance (GIV) method, but they cannot be combined in the same forest plot, as the control arm would be counted more than once. Objectives: To find a simple adjustment that allows the paired comparisons to be combined in the same forest plot by adjusting the standard error (SE). Methods: From first principles it is possible to derive a formula for adjusting the SE in each paired comparison in order to allow pooling in a forest plot. The adjustment is based on the assumption of equal variance in each of the trial arms. The method should not be used if the size of each trial arm is clearly unequal. Results: - If there are N active arms in the study, the SE for each paired comparison should be multiplied by the square root (SQRT) of (N+1)/2)). So for a three-arm trial with two active arms and one control, the SE for each of the two paired comparisons should be multiplied by SQRT(3/2) = 1.225. - Where it is possible to calculate the average of the mean differences (for example) across the paired comparisons, and each paired comparison has a similar 95% CI, then the appropriate SE for the combined comparison groups compared with control can be obtained (under the same assumptions of equal variance across the arms) by dividing the paired SE by SQRT ((N+1)/2N). Conclusions: A simple method is available to allow pooling of data from multi-arm trials in the same forest plot using GIV analysis. However, if a random-effects meta-analysis is being used, combining the active arms will give more weight to the study data than separate paired comparisons. RO 9.8 Meta-analysis of adverse events data using a hierarchical Bayesian model: a case study Mesgarpour B1 , Schmitz S2 , Walsh C3 , Herkner H4 1 National Institute for Medical Research Development (NIMAD), Iran; 2 Department of Pharmacology and Therapeutics, Trinity College Dublin, Ireland; 3 Department of Statistics, Trinity College Dublin, Ireland; 4 Department of Emergency Medicine, Medical University of Vienna, Austria Background: Reporting adverse events of pharmacological treatments is generally inadequate and incomplete in randomized clinical trials (RCTs). The safety profile of medications can be assessed more robustly in the long-term follow-up of observational studies such as cohorts. Standard methods of incorporating both study designs put much weight on the usually large observational studies, despite their, Cochrane Database Syst Rev Suppl 1–327 (2015) 62 presumably, lower internal validity. Objectives: A case study to combine safety data from RCTs and observational studies into a hierarchical Bayesian model adjusting for design-related weights. Methods: We set up a systematic review on adverse events of erythropoiesis stimulating agents (ESAs) in off-label indications in critically ill patients. Eleven databases were searched up to April 2012. We considered RCTs and controlled observational studies in any language that compared off-label ESAs treatment with other effective interventions, placebo or no treatment in critically ill patients. We used frequentist and Bayesian models to combine studies, and performed sensitivity analyses by using different weights for the observational trials. Results: We included 48 studies from 12,888 citations (34 RCTs; 14 observational). In order to combine data from RCTs and observational studies, we fitted a three-level hierarchical Bayesian model, which accounts for between-trial design heterogeneity. ESAs increased the risk for venous thromboembolism (VTE) in the RCTs using frequentist analyses, but had no effect on VTE in the observational studies, or when Bayesian methods were applied. We found no statistically significant difference in the mortality risk from treatment with ESAs compared to non-ESAs in a Bayesian estimate of combining data from RCTs and observational studies. Sensitivity analysis using the Bayesian approach is consistent with the main analysis. Conclusions: A Bayesian approach can be used in a systematic review of adverse events to combine the information from all available controlled study designs by putting less weight on the observational studies. RO 9.9 Prediction intervals should be routinely reported in meta-analyses IntHout J1 , Ioannidis JP2 , Rovers MM1 , Goeman JJ1 1 Radboud University Medical Center, The Netherlands; 2 Stanford University, USA Background: Evaluating the variation in the strength of the effect across studies is a key feature of meta-analyses. This variability is reflected by measures like τ 2 or I2 but their clinical interpretation is not straightforward, especially not if the meta-analysis is in odds ratios or risk ratios. Objectives: A prediction interval presents the expected range of true study effects in similar studies. We aim to show how it can help to understand the uncertainty about whether a treatment works or not. Methods: Conclusions based on confidence intervals (CI) may not hold in all settings. We evaluated the differences in conclusions based on 95% CIs and 95% prediction intervals in statistically significant meta-analyses published in the Cochrane Database of Systematic Reviews between 2009 and 2013. If the estimated I2 was > 0 we used this I2 for the prediction interval. When the estimated I2 was 0, we imputed low levels of heterogeneity (I2 = 20%), because heterogeneity estimates are often imprecise. Results: A statistically significant meta-analysis does not guarantee that the treatment will be effective in all settings: in 347 (72%) of the 479 statistically significant meta-analyses with Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. I2 > 0, the prediction interval showed that the treatment could be ineffective. Imputation of an I2 of 20% in the other 441 statistically significant meta-analyses gave similar results: 329 (75%) of the treatments could be ineffective. Conclusions: The CI is inadequate for clinical decision making because it only summarizes the average treatment effect. The prediction interval is more informative as it shows the range of possible effects in relation to the no-effect and clinical benefit thresholds. A narrow prediction interval completely on the beneficial side of a clinically relevant threshold increases confidence in an intervention. A broad prediction interval may indicate the existence of settings where the treatment has a suboptimal and possibly even harmful effect. Prediction intervals should be routinely reported to allow more informative inferences in meta-analyses. Rapid oral session 10 Filtering the information overload for better decisions RO 10.1 Integrating evidence: existing systematic reviews and primary studies Robinson K1 , Chou R2 , Berkman N3 , Newberry S4 , Fu R5 , Harling L6 , Dryden D6 , Butler M7 , Foisy M8 , Anderson J9 , Motu’apuaka M9 , Revelo R9 , Guise J9 , Chang S10 1 Johns Hopkins University EPC, USA; 2 Pacific Northwest EPC, USA; 3 RTI International, USA; 4 Southern California EPC, RAND, USA; 5 Portland VA Research Foundation, USA; 6 Department of Pediatrics, University of Alberta, Canada; 7 Minnesota EPC, USA; 8 Department of Pediatrics, University of Alberta, USA; 9 Scientific Resource Center for the AHRQ Effective Health Care Program, USA; 10 Center for EvidenceandPracticeImprovment, AgencyforHealthcareResearch and Quality, USA Background: The exponential growth in the number of systematic reviews (SRs) being published, the need to regularly update existing reviews, along with the time and resources required to undertake a review, motivates the desire to integrate existing reviews into a new review. However, there is a trade-off between accepting the results of a prior review and redoing selected elements, or the review in its entirety. Objectives: The goal of this systematic review methods development project was to enhance guidance on approaches and considerations for integrating existing SRs into new reviews. Methods: Over a two-year period, a workgroup of SR methodologists from across the US Agency for Healthcare Research and Quality (AHRQ) Evidence-based Practice Centers (EPCs) clarified guidance needs and developed new recommendations. Deliberations were supported by a review of guidance developed by the EPCs and eight other organizations, including Cochrane and the National Institute for Health and Care Excellence (NICE ); interviews with key informants from organizations Cochrane Database Syst Rev Suppl 1–327 (2015) 63 that conduct or use SRs; and a search of the literature. Recommendations were developed in an iterative manner and based on consensus of workgroup members. Results: In Year 1, the workgroup found general consensus across organizations for assessing the relevance of prior reviews and scanning their references to identify studies for a new review. In Year 2, the workgroup identified eight areas where additional guidance was needed. No literature relevant to informing discussions was identified. Recommendations were developed concerning: minimum eligibility criteria for including an existing review based on its approach and methodological rigor; criteria for using risk of bias assessments from an earlier review; presenting findings from the earlier review distinctly from new studies and synthesizing the two; and an approach for strength of evidence. Conclusions: Preliminary guidance was developed promoting an efficient and unbiased approach for integrating old and new evidence. RO 10.2 Incorporating non-trial data into systematic reviews: opportunities and risks Tsafnat G1 , Gallego-Luxan B1 , Coiera E1 1 Centre for Health Informatics, Australian Institute of Health Innovation, Macquarie University, Australia Background: Clinical data and measures that do not come from clinical trials (e.g. electronic health records, social media) bring additional information that may enhance traditional systematic reviews of the evidence. Current methods for critical appraisal cannot be applied to such data. Emerging methods for interpreting evidence, (e.g. crowdsourcing, automatic systematic reviews [1]), may provide opportunities to incorporate non-trial data whilst ensuring integrity and reliability. Objective: To examine the opportunities and risks of integrating non-trial data into systematic reviews using novel methods and technologies. Results: Non-trial data offer the following advantages: - More data and longer follow-up [2] - Better representation of healthcare services including comorbidities and settings - Include non-drug therapies Data for circumstances where randomized controlled trials are unethical or infeasible - Broader sets of patient outcomes - Capture rare events The following risks also exist: - Lack of randomisation, control and unmeasured confounders introduce biases - Self-reported data (e.g. from social media) may have selection bias - Limited to retrospective analysis and probably unavailable for new therapies These risks may be mitigated by: - Transparent, objective and repeatable protocols in automatic systematic reviews [1] Crowdsourcing editorial services for robust fact-checking Control of bias using multiple datasets in lieu of multiple arms, and statistical methods that ensure significance of results - Monitor conventional and social media for reports on harms or harmful sentiment (e.g. anti-vaccine) that may affect practice [3] to bias but provide crucial information unavailable from clinical trials. More research is needed on when and how non-trial data can be used in systematic reviews. Reference: Tsafnat et al. The automation of systematic reviews. BMJ 2013: 346: f139. 2. Gallego et al. Role of EHR in comparative effectiveness research. JCE 2013: 2: 529–532. 3. Zhou et al. Using social connection information to improve opinion mining. MedInfo 2015. RO 10.3 Application of text mining and machine learning for problem formulation in systematic reviews Thayer K1 , Howard B2 , Holmgren S1 , Pelch K1 , Walker V1 , Lunn R1 , Shah R2 1 National Institute of Environmental Health Sciences (NIEHS)/National Institutes of Health (NIH), USA; 2 SciOme LLC, USA Background: Identifying addressable questions for systematic reviews can be a challenge, especially in environmental health where evidence from human, animal, and in vitro studies is often integrated in assessments. Text-mining and machine learning tools hold promise to help with problem formulation. Objectives: To explore the utility of using the Sciome Workbench for Interactive, Computer-Facilitated Text-mining (SWIFT) software to visualize literature search results for three complex topics: research trends for ∼500 endocrine-disrupting chemicals; environmental influences on the epigenome; and health effects associated with night shift work, light at night, or circadian disruption. Methods: Literature search results from PubMed were uploaded into SWIFT for each project. Customized search strategies were developed for evidence stream (i.e. human, animal, in vitro), exposure, and health outcome. The unsupervised topic clustering functionality of SWIFT was used to group articles by subject matter. Users created intersections of various tags to focus on specific topics, e.g. night shift work and metabolic disorders. Together these functions were used to create interactive reports. Results: The interactive, visual reports produced by SWIFT allowed users to identify and formulate focused research questions more efficiently. The reports helped identify topics that have been extensively studied, as well as emerging areas of research (Figure 1). Viewing results by evidence stream helped users determine how much evidence integration might be required (Figure 2). The topic clustering results are also used to identify ’seed studies’ for the purpose of training a machine-learning model that priority ranks relevant studies in focused areas. Conclusions: Text-mining and machine learning programs such as SWIFT are valuable tools for problem formulation. These types of analyses could be considered for a type of scoping review that can be used for various purposes, ranging from showing trends in research, to identifying targeted questions that could be addressed in systematic reviews. Conclusion: Non trial sources of clinical data are more prone Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Cochrane Database Syst Rev Suppl 1–327 (2015) 64 Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Cochrane Database Syst Rev Suppl 1–327 (2015) 65 RO 10.4 Mining the information overload: a scoping review of patients’ responses to different forms of health risk information Pennington A1 , Noble C1 , Garner J1 , Harris R1 1 University of Liverpool, United Kingdom Background: The explosion of new information technologies and the global growth in the academic sector present challenges to systematic reviewers that have grown in parallel to the burgeoning of electronic databases. Reviewers are increasingly at risk of either being overwhelmed by unmanageable search results, or missing evidence. The challenges are greatest for reviews of ’broad’ social determinants of health and interventions, particularly when they include phrases such as ’health information’. New information technologies/methods also provide potential solutions. Objectives: Our scoping review aimed to describe the extent and nature of evidence on patients’ responses to different forms of information on their health status/risk of disease, with the objectives of mapping key concepts, sources and types of evidence, commonalities, themes and gaps in the research. Methods: To strike the balance between sensitivity (finding all articles in an area) and specificity (finding only relevant articles), the review used standard systematic review approaches to develop search strategies in conjunction with text mining approaches utilising Automatic Term Recognition (ATR) software. Sample papers were identified in initial searches and screened for inclusion. ATR software identified search terms/phrases within the sample papers, producing 107 precise phrases (as opposed to Boolean searches of words, e.g. ’health NEAR/3 information’) after manual selection by reviewers. Phrases were combined with simple operators and run across nine databases. Results: The searches identified 6662 unique articles, compared to standard approaches which identified 100,000+ articles. Title and abstract screening identified 358 papers for full text screening - high numbers here being a measure of the sensitivity of the search of this broad area. Twenty-six articles were included in the review. Conclusions: Our systematic approach combining traditional search term development methods and new text mining methods produced results that were sensitive, specific and, of growing importance to the increasingly overwhelmed community of reviewers, manageable. RO 10.5 Using innovative methods in trial identification for 12 Cochrane Intervention Reviews Noel-Storr A1 , Ware J1 , Gouda P2 , Hull M3 , Jackson D1 , Murrieta Álvarez I4 , Yaman H5 , Forbes S6 , Rutjes A7 1 Cochrane Dementia and Cognitive Improvement Group, Oxford University, UnitedKingdom; 2 NationalUniversityofIreland, Ireland; 3 George Mason University, USA; 4 Universidad Popular Autónoma del Estado de Puebla, Mexico; 5 Akdeniz University, Turkey; 6 Okanagan College, Canada; 7 University of Bern, Switzerland Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Background: In June 2014 the Cochrane Dementia and Cognitive Improvement Group began a programme of work to produce 12 intervention reviews looking at modifiable risk factors (MRFs) for dementia. To produce this large body of evidence the Cochrane Dementia Group have implemented a number of innovative methods in review production. This study focuses on the two innovative methods used in trial identification. Objectives: Our objective was to identify all relevant trials for potential inclusion in 12 intervention reviews related to MRFs in dementia using innovative methods, including crowdsourcing, and to evaluate those methods against appropriate reference standards. Methods: Two main innovations were utilised in the search and screen process. The first was to take a ‘suite’ approach and develop one search strategy per suite of related reviews rather than one search per review; and the second was to recruit a crowd through Students for Best Evidence to screen the search results using an online screening tool. Both methods were to be validated or compared against traditional methods. Results: Taking a suite approach to the search meant that far fewer unique sets of search results were produced (four instead of 12). This meant that fewer overall results were identified with no compromise on the sensitivity of the searches. A total of 48 participants signed up to screen citations, with 23 (48%) screening 500 or more. A deadline of four weeks was given and met. Crowd performance, measured against random samples screened by members of the core teams (under evaluation), will be presented in terms of the crowds’ sensitivity (the collective ability to identify the RCTs correctly) and specificity (identify the rejects correctly). Conclusions: We made gains in efficiency through having fewer citations to screen overall and by harnessing a crowd to screen those citations for trial design. This work contributes to the growing body of evidence on the beneficial role of crowdsourcing in the review production process. As we are confronted with ever increasing amounts of data to process, we need to find new methods to deal with the information overload. RO 10.6 Enhancing search strategy development using word clouds Robalino S1 1 Newcastle University, United Kingdom Background: Systematic reviews increasingly seek to address issues beyond the relative effectiveness of interventions. This can be challenging when the concepts to be searched for are poorly indexed or difficult to describe. Using two case studies, one focusing on the ethics and acceptability of an intervention and the other focusing on economic evaluations, an approach to enhancing search strategy development is outlined. Objectives: To use text mining principles to examine metadata and full-text papers to discover terms and concepts not discovered in standard search strategy development. Methods: Using known relevant papers, metadata from bibliographic databases and full-text papers alongside freely available word cloud generators, both MeSH and free-text terms were examined to highlight undiscovered descriptors Cochrane Database Syst Rev Suppl 1–327 (2015) 66 for use in a systematic review search strategy. Initially text was taken from the bibliographic database metadata (MeSH and keywords), title and abstract, and full-text papers. These sets of text were then examined separately to determine the most useful texts to explore through word clouds which highlight frequency of terms in their display. Results: The use of word clouds to discover ‘hidden’ MeSH and keywords for search strategies and filters works well. In both case studies its use added to the final search strategies in both projects. The inherent simplicity of word clouds makes this an accessible method of discovery, however some types of text yielded better results than others, due to the limitations of word cloud software. Conclusions: This method is easily accessible to anyone wishing to develop search strategies for hard-to-define or unfamiliar topics as it does not require large volumes of texts nor expertise in programming. This method of discovery can be utilised alongside standard approaches to provide a richer set of terms with minimal additional effort. RO 10.7 The Embase project 3: the 48-hour citation screening challenge Noel-Storr A1 , Dooley G2 , Glanville J3 , Foxlee R4 1 Cochrane Dementia and Cognitive Improvement Group, Oxford University, United Kingdom; 2 Metaxis Ltd, United Kingdom; 3 York Health Economics Consortium, United Kingdom; 4 Cochrane Editorial Unit, United Kingdom Background: The Embase project aims to identify all reports of randomised trials and feed those into Cochrane’s Central Register of Controlled Trials (CENTRAL). During 2011, 2012 and 2013 a backlog of records to screen had built up. To tackle the backlog in a time efficient way two methods were employed: (1) a small team of screeners were assigned to focus on these records; and (2) a citation screening challenge was organised. Objectives: The main objective of the citation screening challenge was to screen as many citations as possible within a 48-hour period. Methods: A challenge crowd was recruited in the weeks before the event. Backlog records were loaded into a copy of the online citation screening tool. The 48-hour period began on 31 October and finished on 2 November. The event was a fundraiser, with a micro amount set for each citation collectively screened. The proceeds were to go to Médecins Sans Frontières for the Ebola relief effort. Results: A total of 75 people from 20 countries took part and 20,709 citations were screened in a 48-hour period; 1713 RCTs/q-RCTs were identified and GBP 5177.25 was raised for the Ebola relief effort. The challenge crowd screened over twice as many records in the final 24 hours of the challenge compared to the first 24 hours. An evaluation of challenge crowd accuracy is underway and will be presented. In an after-challenge questionnaire completed by 84% of participants, 95% said they would take part in another challenge. Conclusions: The 48-hour citation screening challenge was the first of its kind and proved an incredibly effective and fun method to help to screen a large number of citations in a very short space of time. Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. RO 10.8 Transforming from information overload to information excitement using the Pearl Harvesting Information Retrieval Framework Sandieson R1 1 Western University, Canada Background: Information overload derives from a number of sources including an expanding body of research that is more dispersed across journals and databases (Hall 2004). Traditional information retrieval strategies are cumbersome in the digital environment (Arendt 2007; Sandieson 2010). The Pearl Harvesting Information Retrieval Framework solution (PHIRF; Sandieson 2006; 2010; 2013) is a design science approach that uses rich-text searches based on synonym clusters. Synonym clusters include ALL the terms used by researchers and indexers to identify a topic. Term harvesting is accomplished from a broad range of sources to avoid the bias of using a restricted set of search terms, which is a potential problem with existing search term development. Present Study: There has been a recent dramatic production and evolution of systematic reviews and research databases have not kept pace with indexing these. Common search strategies to search for systematic reviews use ’systematic review’ OR ’meta analysis’ OR ’literature review’. The term ’review’ is recommended, but has very low precision. In the present study the PHIRF was used to produce a comprehensive synonym cluster for systematic reviews. Results: The Pearl Harvesting systematic review synonym cluster contained 25 terms and produced 57% more citations than the combined standard search terms (excluding ’review’; using .af) in MEDLINE. When paired with a synonym cluster for autism, the PH search produced 1146 citations versus 685 using the standard search. In PsycINFO, the PHIRF search produced 90% more systematic review citations than the standard search (using ALL). The PHIRF search produced 4104 systematic review citations for autism versus 3716 using the standard method. Conclusions: Pearl Harvesting improved the number of citations retrieved for systematic reviews in two sample databases. Synonym clusters, once established, are placed in a public wiki for anyone to use. Our experience of teaching people how to do PHIRF searches is that when they experience comprehensive, relevant results they transform from a state of information overload to information excitement. References: Arendt, J. (2007). How Do Psychology Researchers Find Studies to Include in Meta-Analyses? Behavioral & Social Sciences Librarian, 21, 1–23. Hall, A. & Walton, G. (2004). Information overload within the health care system: a literature review. Health Information and Libraries Journal, 21, 102–108. Sandieson, R. (2006). Pathfinding in the research forest: The Pearl Harvesting method for effective information retrieval., Education and Training in Developmental Disabilities, 41(4), 401–409. Sandieson, R.W., Kirkpatrick, L.C., Sandieson, R.M., Zimmerman, W. (2010). Harnessing the Power of Research Databases: The Pearl Harvesting Methodological Framework Cochrane Database Syst Rev Suppl 1–327 (2015) 67 for Information Retrieval. The Journal of Special Education. 44, 161–175. Sandieson, R.W., & McIsaac, S. M. (2013). Navigating the information maze of giftedness using the Pearl Harvesting Information Retrieval Methodological Framework. Talent Development and Excellence 5(2), 101–112. RO 10.9 Rapid oral session 11 Knowledge translation RO 11.1 Does the medium matter when getting the message? Systematic review on social media as a data source for information on adverse effects Hanratty J1 1 Queens University Belfast, United Kingdom Golder S1 , Norman G2 , Loke Y1 1 Cochrane Adverse Effects Methods Group, United Kingdom; 2 Cochrane Wounds Group, United Kingdom Background: Cochrane Reviews provide high quality summaries of evidence with the healthcare professional in mind, but time constraints often mean that professionals only read summaries or abstracts. Objectives: The aim is to test understanding of the key messages in a review when participants are randomly assigned to receive a summary of a review in one of four mediums (conditions). Methods: Participants were recruited online via mailing lists of health professionals and university staff. A recent Cochrane Review of Parent Infant Psychotherapy (Barlow 2015) was chosen. Participants were randomly assigned to one of four conditions; reading the review abstract, the plain language summary, a blog post aimed at health professionals or listening to a podcast interview with the review author. Participants were then asked about the key findings of the review and how, if at all, it would impact on their practice. Results: This study is ongoing and results will be presented in terms of the proportion of respondents in each condition who correctly identified the core messages of the review. Background: Social media (such as patient forums, Twitter, and Facebook) are increasingly popular and contain a vast array of unpublished up-to-date information. There are two avenues through which social media can provide adverse effects data: first by providing references to published and unpublished literature and second by providing patients’ experience of adverse effects. Social media have not yet been fully explored as a potential source of adverse effects data for systematic reviewers. Objectives: To review the literature that has evaluated systematically or analyse the use of social media to collect information on adverse effects. Methods: Eighteen databases (including MEDLINE and Embase) were searched for relevant studies, in addition to handsearching key journals, conferences, newsletters and blogs, contacting experts and reference checking. Any type of evaluation was considered eligible for inclusion if it assessed the use of social media to collect information on adverse effects and presented the outputs from searching social media. Data extraction and quality assessment was undertaken independently by two reviewers. Due to the nature of the included studies the quality criteria were customized to evaluate aspects such as the selection of data, duplicate data and misinformation and validity and representativeness. Results: A total of 3045 records were retrieved (4457 before duplication). The included studies were heterogeneous in nature and thus a narrative synthesis with descriptive analysis was undertaken. We summarise the existing research on the potential value of social media for information on adverse effects and discuss the implications for systematic reviewers. Conclusions: Case reports of adverse effects identified in social media could be a useful source of evidence in systematic reviews. In addition, posts on social media could help us identify those adverse effects most important to patients and thus help us formulate and prioritise questions on adverse effects for future systematic reviews. Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. RO 11.2 Communication of systematic review findings: exploring format preference Thompson Coon J1 , Bethel A1 , Abbott R1 , Rogers M1 , Moore D1 , Stein K1 1 University of Exeter Medical School, United Kingdom Background: Ensuring that research findings reach their intended audience so that they may make a difference to health care is a challenge that is gaining greater emphasis. Traditional methods of dissemination may result in limited communication. Method: Following the completion of a systematic review about the use of gardens by people with dementia resident in care homes, we produced a range of dissemination products containing the findings of the review: a peer-reviewed academic paper, an A4 hand-out of a conference poster, a tri-fold leaflet and a video clip. At a multi-disciplinary dementia conference we asked individuals to choose which format they preferred and to give their reasons for this choice. Results: The activity attracted a lot of attention from conference attendees who were keen to participate and help to identify approaches to make research findings more readily accessible to them. A total of 40 people took part; 16 health care professionals (including nurses, occupational Cochrane Database Syst Rev Suppl 1–327 (2015) 68 therapists and social workers), four carers/members of the public, four lecturer/researchers, two policy makers, four students (social work, nursing), three trainers of healthcare professionals, three members of other support organisations (housing, memory café) and three who did not specify. The leaflet was the most popular option (18 votes), followed by the video (9 votes), the poster (8 votes) and the peer-reviewed paper (5 votes). The most common reasons for preferring the leaflet were that it was easy to read and easy to access. Conclusions: These pilot results suggest that consumers of systematic review findings are seeking to engage with them in a variety of different formats. Ease of access and the use of lay language are important considerations. Further work is planned to explore preferred level of content, the layout of dissemination products and methods of distribution. RO 11.3 Adapting fishbone diagrams: a new approach for displaying findings of systematic reviews Morgan L1 , West S1 , Gartlehner G2 , Jordan H1 , Kampov-Polevoy A3 , Garbutt J3 , Bobashev G1 1 RTI International, USA; 2 DUK, Austria; 3 University of North Carolina, Chapel Hill, USA Background: The traditional approach to presenting the results of systematic reviews is to use multiple summary tables. Although highly informative, these tables are often dense and do not allow a reader to evaluate the evidence on various and outcomes ‘‘at a glance’’ and instead require large parts of the publication to be read. Since they can easily and clearly illustrate cause and effect, fishbone diagrams are widely used in industrial studies and are beginning to be used in health care. The goal of this paper is to show how these diagrams can be adapted for use in graphically capturing the results of systematic reviews in health care research, and to stimulate further consideration of pictorial summary approaches. Methods: We applied the fishbone diagram to summarize the findings from a sample of systematic reviews visually. The head of the fish represents the balance of benefits and harms of an intervention or comparison, and the bones of the fish represent the individual health outcomes or moderators. To illustrate a range of applications, we abstracted and compiled information from the summary tables of three recent systematic reviews to populate each fishbone diagram. Results: Based on real world systematic reviews, we provide three examples that show how fishbone diagrams might be used to display complex information in a clear and succinct manner. We present fishbone diagrams depicting: 1. the body of evidence used for a clinical practice guideline; 2. the body of evidence for development of population-based screening recommendations; and 3. the relationship between an outcome and various positive and negative moderators. Conclusions: Fishbone diagrams are Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. compact visualizations that may prove useful for summarizing the findings of systematic reviews, especially for healthcare providers and guideline developers who do not have the time to review the full evidence in its entirety. RO 11.4 Interactive ’Summary of findings’ table (iSoF) Rosenbaum S1 , Morelli A1 , Moberg J1 , Oxman A1 , Rada G2 , Collaboration D3 1 Norwegian Knowledge Centre for the Health Services, Norway; 2 Evidence-Based Health Care Program, Pontificia Universidad Católica de Chile, Chile; 3 GRADE Working Group, EU Background: When making informed healthcare decisions, people need to understand the likely risks and benefits of the treatments they are considering. ’Summary of findings’ (SoF) tables help people understand key messages from Cochrane Reviews in a concise format. However, many people still have difficulty understanding the numbers and some of the concepts in SoF tables. Static SoF tables also lack formatting flexibility and limit the amount of detail that can be presented to more expert users. Objective: To create an interactive SoF table (iSoF) that helps all users find and understand the main messages from a systematic review by allowing them to choose the amount of detail and mode of presentation that is appropriate for them to inform their decision-making. Method: The iSoF was developed iteratively by the DECIDE project (funded by the EU) informed by feedback from stakeholders, user testing and an advisory board of international experts. Results: iSoF tables present the key messages from an evidence summary: the most important outcomes (benefits and harms), the size of the effects, and the certainty of the evidence. They offer a layered presentation, enabling producers to tailor tables for different audiences or types of evidence or users to ‘drill-down’ for more information by scrolling over terms, concepts or interactive footnotes to find explanations. Each outcome can be viewed in several formats: plain language that describes effect sizes and the certainty of the evidence in a way most people can readily understand; absolute effects presented as numbers, text, or interactive graphic visualizations; and as relative effects. iSoF tables provide an optimal starting point for other evidence summaries of a review, as they contain all the building blocks necessary to present key findings in concise, understandable format. iSoF tables can be prepared in different languages and exported for use in other documents. Conclusion: iSoF tables give producers and end users with varying degrees of expertise flexibility, control and support for understanding the main findings of an evidence summary. http://isof.epistemonikos.org/#finding/5263ba01f30d0c11a5 3f7950 Cochrane Database Syst Rev Suppl 1–327 (2015) 69 RO 11.5 RO 11.6 Can an evidence-informed taxonomy be used to map vaccination communication interventions in ‘real world’ settings? Findings from Nigeria Glossary of Evaluation Terms for Informed Treatment choices in plain language (GET-IT) Oku A1 , Oyo-ita A1 , Lewin S2 , Glenton C2 , Fretheim A2 , Ames H2 , Kaufmann J3 , Hill S3 , Cartier Y4 , Cliff J5 , Muloliwa A6 , Rada G7 , Bosch-Capblanch X8 1 University of Calabar, Nigeria; 2 The Norwegian Knowledge Centre for the Health Services, Norway; 3 La Trobe University, Australia., Australia; 4 International Union for Health Promotion and Education, France; 5 Universidade Eduardo Mondlane, Mozambique; 6 Provincial Directorate of Health, Mozambique; 7 Catholic University of Chile, Chile; 8 Swiss Tropical and Public Health Institute and University of Basel, Switzerland Background: Effective communication is key to improving childhood vaccination coverage and is the focus of the Communicate to Vaccinate (COMMVAC) project. In earlier work, we developed a taxonomy of communication interventions to help decision makers understand and consider options for communication. As interventions in systematic reviews become more complex and multifaceted, such taxonomies are also assuming greater relevance for systematic review methods. Objectives: To describe the application in the Nigerian context of a taxonomy of vaccination communication interventions and to explore the usefulness of the taxonomy for grouping and describing interventions and identifying important gaps. Methods: The study was conducted in Bauchi and Cross River States in North and Southeast Nigeria respectively. Interviews were carried out amongst purposively selected stakeholders in the health services and in agencies involved in vaccination communication planning. Document review was also conducted. The COMMVAC taxonomy of vaccination communication interventions was used to organize the identified strategies based on their intended communication purpose and their target group. Results: Most of the communication strategies identified fell into the COMMVAC taxonomy categories of ‘inform and educate’, targeting caregivers and community members, and ‘enhancing community ownership’. Religious and traditional leaders, community mobilizers and women’s groups were common targets for immunization messages and for efforts to sustain community demand in both Bauchi and Cross River, though more visibly in Bauchi. Most communication interventions were used in the context of vaccination campaigns rather than routine immunization programmes, and interventions targeting health workers were also limited. Conclusions: The evidence-informed taxonomy was helpful in mapping communication strategies for childhood vaccination. It could assist programme managers to identify gaps and consider the communication purposes of the interventions they are using. The study suggests that evidence-informed taxonomies may be a useful tool for mapping interventions in ’real world’ settings. Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Moberg J1 , Austvoll-Dahlgren A1 , Treweek S2 , Badenoch D3 , Harbour R4 , Rosenbaum S1 , Oxman A1 , Chalmers I5 1 Norwegian Knowledge Centre for the Health Services, Norway; 2 Health Services Research Unit, University of Aberdeen, United Kingdom; 3 Minervation, United Kingdom; 4 Glasgow, United Kingdom; 5 James Lind Initiative, United Kingdom Background: Well-informed healthcare decisions depend on the ability to understand information, particularly research evidence. The use of jargon can be a barrier to people’s understanding and use of evidence to inform their choices. Objective: To create an online glossary that facilitates informed choices about treatments by promoting consistent use of plain language, providing plain language definitions and explanations of terms that people might need to understand if they wish to assess claims about treatments. Methods: GET-IT (Glossary of Evaluation Terms for Informed Treatment choices) was developed collaboratively by the Informed Healthcare Choices project (funded by GLOBVac), Testing Treatments interactive, and the DECIDE project (funded by the EU). We screened 15 sources to identify terms, including a sample of relevant glossaries. Evaluation of the glossary includes feedback from collaborating partners, a survey of organisations that have expressed an interest in using GET-IT, user testing, and an evaluation of the impact of the glossary on users’ understanding of terms. Results: The glossary includes over 200 terms. For each term, there is a suggested plain language term, synonyms, a short definition, a full explanation, examples, and links to additional resources. The glossary can be translated, tailored to different audiences, and embedded in other websites. Other websites can also link to definitions. Conclusions: The GET-IT glossary is open-access, and the use of plain language facilitates understanding by a wide range of people. It is useful for people who communicate research evidence (including systematic reviewers), and who teach others about how to assess claims about treatments, in addition to those who want to make informed choices about treatments. RO 11.7 A review of online evidence-based practice point-of-care information summary providers: 2014 update Kwag KH1 , Gonzàlez-Lorenzo M2 , Banzi R3 , Bonovas S4 , Moja L1 1 Clinical Epidemiology Unit, IRCCS Galeazzi Orthopaedic Institute, Milan, Italy; 2 Clinical Epidemiology Unit, IRCCS Galeazzi OrthopaedicInstitute, Milan; DiparmentofBiomedichalSciencesfor Health, UniversityofMilan, Italy; 3 LaboratoryofRegulatoryPolicies, IRCCS-Institute for Pharmacological Research Mario Negri, Milan, Italy; 4 Department of Biomedical Sciences for Health, University of Milan, Italy Cochrane Database Syst Rev Suppl 1–327 (2015) 70 Background: The complexity of modern practice requires health professionals to be active information-seekers. Evidence-based point-of-care (EBP) information summaries can connect clinicians with best practice strategies at the bedside to optimize patient care. This study is the 2014 update of a previous review that assessed the strength of these products. Objective: To review EBP summaries against their claims of being ‘‘evidence-based,’’ and determine their progress. Methods: We searched MEDLINE, Google, librarian association websites, and information conference proceedings from August 2012 to December 2014. We included online EBP summaries delivering literature syntheses that claim to be evidence-based. We extracted data on general characteristics and content presentation. We assessed products quantitatively according to breadth, editorial quality, and evidence-based methodology. We explored improvements in these dimensions and their relevant associations. Results: We screened 58 online products: 26 met our inclusion criteria, 10 of which were newly identified in 2014. We were able to access 23 products. Most summaries were produced by major publishers in the USA, with a minority in Europe and none in developing- or low-income countries. No significant association between the pairs of variables was found. There were improvements in evidence-based methodology and breadth since 2009. The main target audience remained physicians, although several products targeted nurses, physiotherapists, emergency specialists, and pediatricians. Conclusions: New EBP summaries are targeting specialized groups of health professionals, reflecting the strength of the existing market of point-of-care summaries for physicians. UptoDate, Best Practice, and Dynamed scored the highest across all dimensions, although other products excelled in one or two dimensions. Consumers should weigh their needs carefully when selecting a product for implementation. New generation EBP summaries aim to fulfill patient-specific needs by integration into Electronic Health Records, as well as providing an option for Continuing Medical Education to promote lifelong, self-directed learning. RO 11.8 Abstracts’ concept selection to reduce information overload Alper B1 , Malone-Moses M1 1 Evidence-Based Medicine Research and Development, Quality and Standards for EBSCO Health, USA Many users will access abstracts of Cochrane reviews to determine key points, either to represent the complete concept or to determine if full-text reading of the review is warranted. Abstracts need to be concise yet include the most important concepts. We will analyze 100 Cochrane review abstracts compared to key concepts selected from the Cochrane reviews for point-of-care summaries. RO 11.9 Development and validation of a questionnaire to measure people’s ability to assess claims about treatment effects (CLAIM) Austvoll-Dahlgren A1 , Oxman AD1 , Chalmers I2 , Semakula D3 , Nsangi A3 , Guttersrud Ø4 1 Norwegian Knowledge Center for the Health Services, Norway; 2 James Lind Initiative, United Kingdom; 3 Makerere University College of Health Sciences, Uganda; 4 University of Oslo, Norway Objective: People are confronted with claims about the effects of treatments and health policies daily. There is little current evidence available that describes people’s ability to assess claims about treatment effects and there is no instrument available to measure this. We are developing such an instrument as part of an international Project - ’Informed Healthcare Choices’ - with the aim of improving the ability of people in low-income countries to assess claims about treatment effects. Methods: We developed multiple-choice items to address 31 key concepts that people need to understand to be able to assess claims about treatment effects. We obtained feedback on the applicability and relevance of the items from an advisory group and conducted cognitive interviews and a field test with children and adults in Uganda. Results: We have prepared a pool of approximately 240 items (six to eight per concept) intended to measure basic and applied understanding using plain language and universally relevant scenarios. Members of the advisory group judged the items to be relevant to the concepts. Terminology, instructions and formats were rewritten to address the feedback from children and adults. The items’ psychometric properties will be assessed using Rasch analysis. The final instrument will include approximately 31 items (one for each of the 31 concepts). Conclusion: The ’CLAIM’ Instrument is, to our knowledge, the first instrument to be developed to measure people’s ability to assess claims about treatment effects. We will use the instrument in randomized trials to measure the effects of the ’Informed Healthcare Choices’ teaching resources and news service. The items will also be posted on the Informed Healthcare Choices website as a teaching resource for people to test themselves Rapid oral session 12 Specific study types / Avoiding research waste RO 12.1 Should case series be included in systematic reviews when controlled trials are scarce? Sjögren P1 , Kindblom J1 , Liljegren A1 , Strandell A1 , Wikberg Adania U1 , Jivegård L1 1 HTA-centrum, Region Västra Götaland, Sweden Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Cochrane Database Syst Rev Suppl 1–327 (2015) 71 Background: Randomized (RCT) or non-randomized controlled trials are regularly used for assessment of certainty of evidence (GRADE) in systematic reviews (SRs) while case series are mostly ignored. For some interventions RCTs are difficult to design, but well-designed case series are available. Objectives: Studying the influence on conclusion and certainty of evidence (GRADE) by using case series in an SR with only one small RCT. Method: A systematic literature search in an SR of hypoglossal nerve stimulation (HGNS) in adults with obstructive sleep apnoea (OSA) not tolerating Continuous Positive Airway Pressure treatment resulted in one small RCT (therapy withdrawal design) and seven case series. The case series (n=232) were assessed by a modified checklist (Guo 2013). For the outcome ‘resolution of OSA’, measured by apnoea-hypopnea index (AHI), the conclusion and the certainty of evidence (GRADE) were assessed separately for the RCT and the case-series and then combined. Results: The RCT evaluated the effect of HGNS switched ‘on’ or ‘off’ during one week, 12 months after implantation, in responders. The intergroup difference in AHI was 11-fold in favour of HGNS ‘on’. The RCT (starting at + + ++) was downgraded two steps for study limitations regarding blinding and uncertain directness and precision, resulting in ++OO. The case series (starting at +OOO), several well-designed, were downgraded two steps for some limitations in blinding, inconsistency, and serious indirectness. Then the case series were upgraded one step for > 2-fold AHI reduction at one year, clearly different from the published natural course (AHI increase over time) of OSA, and an additional step due to large upper airway volume increase with immediate resolution of OSA with HGNS, resulting in +OOO. Combined, the certainty of evidence was low (++OO). Conclusion: Inclusion of case series reduced the estimate of the treatment effect on AHI from 11-fold in the RCT (short term) to 2-fold in long-term, while the certainty of evidence remained at ++OO. Case series may provide additional information in an SR when controlled studies are scarce and the natural course of the condition is known. Reference: Guo B, Moga C, Schopflocher D, Harstall C. Validation of a quality assessment checklist for case series studies. In: Better Knowledge for Better Health. Abstracts of the 21st Cochrane Colloquium; 2013 19–23 Sep; Québec City, Canada. John Wiley & Sons; 2013 RO 12.2 Integration of multiple study designs in systematic reviews Peinemann F1 , Tushabe D2 , Kleijnen J1 1 Maastricht University, The Netherlands; Birmingham, United Kingdom 2 University of Background: A systematic review may evaluate different aspects of a healthcare intervention. To accommodate the evaluation of various research questions, the inclusion of more than one study design may be necessary. Objectives: To find and describe articles on methodological issues concerning the incorporation of multiple types of study designs in systematic Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. reviews on healthcare interventions. Methods: We searched PubMed, the Cochrane Database of Systematic Reviews, and the Cochrane Methodology Register on 31 March 2012 and identified 42 articles that reported on the integration of single or multiple study designs in systematic reviews. We summarized the contents of the articles qualitatively and assessed theoretical and empirical evidence. Results: Many examples of reviews incorporating multiple types of studies exist and every study design can serve a specific purpose. 85% (36 of 42) of reviews reported that nonrandomized studies should be integrated in systematic reviews to complement available RCTs or replace lacking RCTs. The clinical questions of a systematic review determine the types of designs that are necessary or sufficient to provide the best possible answers. Conclusions: The integration of multiple study designs in systematic reviews is required if patients are to be informed about the many facets of patient-relevant issues of healthcare interventions. RO 12.3 There is no such thing as a one-dimensional hierarchy of evidence: a critique and a perspective Hannes K1 , Bath-Hextall F2 , Behrens J3 1 Cochrane Qualitative and Implementation Methods Group, 2 University of Nottingham, United Kingdom; Belgium; 3 Martin-Luther-Universität, Germany There are many reasons why people favor particular study designs in the context of producing systematic reviews. In a Cochrane context, this choice has historically been guided by the type of design that was most likely to produce robust research findings: the randomized controlled trial, recently extended with a theoretical option of high quality observational research without confounders (GRADE: Grading of Recommendations Assessment, Development and Evaluation). It was argued that decisions toward patients and clients should primarily be based on such evidence. In this paper we argue that this is a destructive choice for two main reasons: (1) it would prevent us from acting in a situation or event for which no robust evidence is available; and, (2) it would lead to an undesirable hierarchy of research questions, with more complex questions not being answered (or funded). We propose two models to respond to these issues: (1) a model integrating external (research driven) and internal evidence (client driven) in supporting evidence-based decision making; and, (2) a question-driven model for a mega-synthesis approach to support review authors in tackling a broad variety of questions. If we are serious about improving health care practice at the point of care, then we should investigate the conditions for realizing a coherent decision making process, based on the evidence that is available (not unavailable) to us. Cochrane Database Syst Rev Suppl 1–327 (2015) 72 RO 12.4 Meta-analyses neglect previous systematic reviews and meta-analyses about the same topic Helfer B1 , Prosser A2 , Samara M1 , Geddes J3 , Cipriani A3 , Davis J4 , Mavridis D5 , Salanti G5 , Leucht S1 1 Cochrane Schizophrenia Group, Germany; 2 Centre for Addiction and Mental Health, Toronto, Canada; 3 University of Oxford, United Kingdom; 4 Cochrane Schizophrenia Group, USA; 5 Cochrane Statistical Method Group, Greece The following abstract is a part of an article accepted for publication in BMC Medicine. Background: As the number of systematic reviews is growing rapidly, we investigate systematically whether meta-analyses published in leading medical journals present an outline of available evidence by referring to previous meta-analyses and systematic reviews. Methods: We searched PubMed for recent meta-analyses of pharmacological treatments published in high impact factor journals. Previous systematic reviews and meta-analyses were identified with electronic searches of keywords and by searching reference sections. We analyzed the number of meta-analyses and systematic reviews that were cited, described and discussed in each recent meta-analysis. Moreover, we investigated publication characteristics that potentially influenced the referencing practices. Results: We identified 52 recent meta-analyses and 242 previous meta-analyses on the same topics. Of these, 66% of identified previous meta-analyses were cited, 36% described, and only 20% discussed by recent meta-analyses. The probability of citing a previous meta-analysis was positively associated with its publication in a journal with a higher impact factor (odds ratio (OR) 1.49; 95% confidence interval (CI) 1.06 to 2.10) and more recent publication year (OR 1.19; 95% CI 1.034 to 1.37). Additionally, the probability of a previous study being described by the recent meta-analysis was inversely associated with the concordance of results (OR 0.38; 95% CI 0.17 to 0.88), and the probability of being discussed was increased for previous studies that employed meta-analytic methods (OR 32.36; 95% CI 2.00 to 522.85). Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Conclusions: Meta-analyses on pharmacological treatments do not consistently refer to or discuss findings of previous meta-analyses on the same topic. Such neglect can lead to research waste and be confusing for readers. Journals should make the discussion of related meta-analyses mandatory. RO 12.5 Unnecessary repetitions of pediatric clinical trials: cumulative meta-analyses Janiaud P1 , Cornu C2 , Kassai B2 1 University of Lyon, France; 2 EPICIME-Clinical Investigation Center, INSERM, France Background: As children are a vulnerable population, it is essential to avoid unnecessary trials when enough evidence is available. Objectives: Our objective was to identify unnecessary repetitions of randomized clinical trials (RCTs) in the pediatric population, using cumulative meta-analyses, when either the beneficial or deleterious effect of the treatment has been confirmed. Methods: We searched three electronic databases for meta-analyses. Double-blind, placebo–controlled RCTs including adults and children were eligible. The random-effects model was used to pool effect over time by including studies according to their year of publication. We performed cumulative meta-analysis to identify significant benefit or risk when multiple RCTs were available for children. Results: Initially 89 meta-analyses were identified, including 992 RCTs amongst which only 294 were in children and evaluating 124 drugs. For 53 drugs, only one pediatric trial was included. Ten drugs for the same indication and with the same outcome were evaluated by more than one meta-analysis. Ondansetron for postoperative nausea and vomiting was assessed by 27 pediatric RCTs. The treatment benefit was confirmed in 1995 (Figure 1). Cumulative meta-analysis showed that 19 trials, which included 840 children, were unnecessary. Conclusions: Our preliminary results suggest the presence in the literature of reports of unnecessary RCTs in children. Clinical research in children should focus on unmet needs and avoid exposing children to unnecessary trials. Cochrane Database Syst Rev Suppl 1–327 (2015) 73 Figure 1 RO 12.6 Does access to clinical study reports from the European Medicines Agency reduce reporting bias? Tonia T1 , Rohner E1 , Petavy F2 , Pignatti F2 , Bohlius J1 1 Institute of Social and Preventive Medicine, University of Bern, Switzerland; 2 European Medicines Agency, United Kingdom Background: In a previous Cochrane Review on the effects of erythropoiesis stimulating agents (ESAs) in cancer patients (1) we identified potential publication and outcome reporting biases for several outcomes. Objectives: We collaborated with European Medicines Agency (EMA) to retrieve clinical study reports (CSRs) and investigated whether inclusion of data reported in these CSRs may help to reduce reporting bias. Methods: We identified all randomised controlled trials (RCTs) on the effect of ESAs on cancer patients from our previous Cochrane Review (1) and updated literature searches. We set up a collaborative agreement with EMA and requested the CSRs for all identified RCTs. We conducted random-effects meta-analyses and compared the pooled estimates based on data as reported in the public domain Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. (i.e. articles, abstracts, other public sources) with the pooled estimates based on data from the public domain and the CSRs. We present preliminary results for the outcomes: number of patients receiving red blood cell transfusions (RBCTs); anaemia symptoms measured with the Functional Assessment of Cancer Therapy-Anaemia (FACT-An); and hypertension; analyses for additional outcomes are on-going. Results: From the review (1) and updated literature searches we identified 92 RCTs, the EMA identified two additional, unpublished studies. For these 94 RCTs the EMA had the CSRs for 16 RCTS. The number of studies and participants analysed and the pooled effect estimates per outcome are shown in Table 1. For RBCTs and hypertension the inclusion of two and nine additional studies, respectively, based on data reported in CSRs did not alter the overall effect estimates. For FACT-An the inclusion of six additional studies based on data reported in CSRs reduced the pooled mean difference of changes from 5.93 (95% CI 4.37, 7.49) to 2.79 (95% CI 0.88, 4.70), which is below the threshold (defined as ≥ 4) of a clinically important difference for FACT-An (1). This result will change the conclusion of the review (1). Conclusions: Unpublished clinical study reports held by EMA may be a useful source to reduce outcome reporting bias. Reference: (1) Tonia et al CD003407. Cochrane Database Syst Rev Suppl 1–327 (2015) 74 RO 12.7 Agreements on publication rights: an investigation of protocols and publications of randomized clinical trials Kasenda B1 , Amstutz A2 , Von Elm E3 , You J4 , Blümle A5 , Tomonaga Y6 , Saccilotto R2 , Bengough T7 , Meerpohl J5 , Stegert M2 , Olu K2 , Tikkinen K4 , Neumann I4 , Carrasco-Labra A4 , Faulhaber M4 , Mulla S4 , Mertz D4 , Akl E4 , Bassler D8 , Busse J4 , Ferreira-González I9 , Lamontagne F10 , Nordmann A2 , Gloy V2 , Raatz H2 , Moja L11 , Ebrahim S4 , Schandelmaier S2 , Sun X12 , Vandvik P13 , Johnston B4 , Walter M14 , Burnand B3 , Schwenkglenks M6 , Kasenda L2 , Bucher H2 , Guyatt G4 , Briel M2 1 Royal Marsden Hospital, United Kingdom; 2 Basel Institute for ClinicalEpidemiologyandBiostatistics, UniversityHospitalofBasel, Basel, Switzerland; 3 Cochrane Switzerland, Institute of Social and Preventive Medicine (IUMSP), Lausanne University Hospital, Lausanne, Switzerland; 4 Department of Clinical Epidemiology and Biostatistics, McMaster University, Hamilton, Ontario, Canada; 5 German Cochrane Centre, Medical Center –University of Freiburg, Freiburg, Germany; 6 Epidemiology, Biostatistics and Prevention Institute (EBPI), University of Zurich, Zurich, Switzerland; 7 Austrian Federal Institute for Health Care, Department of Health and Society, Vienna, Austria; 8 Department of Neonatology, University Hospital Zurich, Zurich, Switzerland; 9 Epidemiology Unit, Department of Cardiology, Vall d’Hebron Hospital and CIBER de Epidemiologı́ a y Salud Publica (CIBERESP), Barcelona, Spain; 10 Centre de Recherche Clinique Étienne-Le Bel and Department of Medicine, Université de Sherbrooke, Sherbrooke, Canada; 11 CS Orthopedic Institute Galeazzi, Milano, Italy; 12 Chinese Evidence-based Medicine Center, West China Hospital, Sichuan University, Chengdu, China; 13 Department of Medicine, Innlandet Hospital Trust-Division Gjøvik, Oppland, Norway; 14 Institute of Nuclear Medicine, University Hospital Bern, Bern, Switzerland Background: For-profit companies conducting randomized clinical trials (RCTs) often make contracts with academic Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. investigators. Little is known about the content of publication agreements documented in RCT protocols, and the accuracy of corresponding statements in published RCTs. Objectives: To investigate: 1. existence and type of publication agreements in RCT protocols; 2. the completeness and accuracy of the corresponding reporting in subsequent publications. Methods: We identified a retrospective cohort of RCTs based on archived protocols approved by six research ethics committees in Switzerland, Germany, and Canada between 2000 and 2003. Last follow-up of RCTs was 27 April 2013. Only RCTs with industry involvement were eligible. Outcomes: Documentation of publication agreements in RCT protocols and their concordance with reported statements in journal publications. Results: Six-hundred and forty-seven RCT protocols were eligible (Figure 1), of these, 456 (70.5%) mentioned an agreement regarding publication of results. Of these, 156 (34.2%) documented an industry-sponsor’s right to disapprove proposed manuscripts; 237 (52.0%) an industry-sponsor’s right to review any manuscript before publication; and 39 (8.6%) of agreements had no constraints regarding publication. The remaining protocols referred to separate agreements that were not accessible to us. We identified 388 publications corresponding to the 647 protocols (60%); in 290 (74.7%) authors did not report any agreement regarding publication although 197 of the corresponding 290 protocols (67.9%) explicitly mentioned such agreements. Only 30 (30.6%) out of 98 (25.3%) publications reporting on agreements about publication were concordant with statements in the corresponding protocols (Table 1). Conclusions: Agreements on publication rights between industry sponsors and academic investigators are common in RCT protocols. In a third, the industry sponsors retain the right to disapprove manuscripts for publication. Journal articles seldom report on publication agreements and, if they do, statements are often discrepant with the trial protocol. Cochrane Database Syst Rev Suppl 1–327 (2015) 75 Figure 1 Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Cochrane Database Syst Rev Suppl 1–327 (2015) 76 RO 12.8 Are Health Technology Assessment (HTA) funded trials using systematic reviews to inform their design? A retrospective cohort Bhurke S1 , Cook A2 , Tallant A1 , Young A1 , Williams E1 , Raftery J3 1 National Institute for Health Research Evaluation, Trials and Studies Co-ordinating Centre (NETSCC), United Kingdom; 2 University of Southampton and University Hospital Southampton NHS Foundation Trusts, United Kingdom; 3 Wessex Institute, University of Southampton, United Kingdom Background: Limited evidence exists on how systematic reviews are used in the design of new trials. Jones and colleagues (2013) investigated this using a cohort of randomised controlled trials (RCTs) funded by the National Institute for Health Research (NIHR) Health Technology Assessment (HTA) programme during the period 2006 to 2008. Their study found that 11 (23%) of 48 applications made no reference to a systematic review. Twenty (54%) of the 37 applications referenced a systematic review and reported their use in designing of the proposed trial. Objectives: To replicate and validate Jones’ study, to explore the reasons for applications not referencing a systematic review. The study also investigated a cohort of NIHR HTA trials funded during 2013 to identify if there were improvements over time. Methods: Two cohorts of NIHR HTA-funded RCTs were included. Cohort I included the same trials as Jones et al (except for one trial that has been discontinued) and cohort II included all trials funded during 2013. Two reviewers undertook data extraction independently. Descriptive statistics were used and no formal statistical comparisons were conducted. Results: Nine trials were identified in cohort I (19%) and three in cohort II (9%) that did not reference a systematic review, but each had a justifiable reason for this. Systematic reviews were referenced in 85% of NIHR HTA trials and nearly 62% referenced more than one systematic review. In total 108 systematic reviews were referenced, 43 of which were Cochrane systematic reviews. Conclusions: Systematic reviews were referenced in 85% of NIHR HTA trials. Fifteen per cent of trials that did not reference a systematic review had justifiable reasons. NIHR requires that proposals for primary research are justified by existing evidence and our study confirms that this was true for all NIHR HTA trials that could do so. RO 12.9 Global evidence, local research: a country map of included studies from Cochrane Systematic Reviews Yao L1 , Cui R1 , Wang X1 , Wang Q1 , Wei D1 1 Evidence-Based Medicine Center, School of Basic Medical Sciences, Lanzhou University, China Background: Cochrane Systematic Reviews (SRs) are one source of high quality evidence, which are widely used Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. in many countries in the development of clinical practice guidelines and health policies. However, worldwide, different ethnic groups have different tolerance levels to the same intervention. So it is important to analyze the geographical distribution of included studies from Cochrane SRs. Objectives: To analyze the geographical distribution of included studies from Cochrane SRs. Methods: Three-hundred Cochrane SRs with full text were randomly selected by computer from the Cochrane Database of Systematic Reviews (CDSR), and the countries or regions of their included studies were analyzed. Results: The 300 Cochrane SRs included 2456 original studies. Among these 2456 studies, 1033 (42%) were conducted in Europe, 779 (32%) were conducted in North America, 442 (18%) were conducted in Asia, 107 (4%) were conducted in Oceania, 60 (3%) were conducted in Africa and 35 (1%) were conducted in South America. Conclusions: Seventy-five per cent of studies were conducted in Europe and North America. Few original studies come from Asia, South America, Africa and Oceania. The unbalanced geographical distribution of original studies might cause indirectness when the Cochrane SRs are disseminated throughout the world. Oral session 7 Communicating evidence O 7.1 Presenting evidence of effectiveness of interventions and more: the evidence of effects page Nunan D1 , Heneghan C1 , Mahtani KR1 , Howick J1 , Thompson M1 1 University of Oxford, United Kingdom Background: Patients are increasingly expected to be actively involved in their care. Moreover, making evidence-based decisions with patients to aid healthcare and treatment decisions is well recognised. Few evidence-based communication tools exist that successfully engage and aid consumer understanding in a format that also facilitates evidence-based clinical decisions. Objectives: To develop a tool that presents the best available evidence for treatment effects that enables both patients and clinicians to make better informed treatment and healthcare decisions. Methods: Using treatment of hypertension with angiotensin converting enzyme inhibitors (ACEi) as an example, we searched the output of the Cochrane Hypertension Review Group for relevant systematic reviews and found one suitable review. We extracted data into Excel on the mean effect (95% confidence limits (Cl)) of 14 ACEi on systolic blood pressure (SBP) only. These data were used to create a modified bar chart. Each ACEi was displayed in descending order according to the certainty of effects based on 95% Cl. The modified chart was inserted into a table column with the heading ’Blood pressure effect performance’. Other columns included ‘Dose’, ‘Cost’, ‘Sample size’ and ‘Duration’. These columns provide data on Cochrane Database Syst Rev Suppl 1–327 (2015) 77 the dosage of drug for the observed blood pressure effect, the cost on a daily basis based on the dose, the number of studies and participants for the observed effect and the duration of treatment/follow-up of these studies. Individual sections with the headings ‘Technical Information’, ‘Cost information’, ‘Dose information’, ‘Quality information’ and ‘Usage information’ are placed under the table and provide clarification and further details of the information contained therein. Results: We present the methods used to derive the first Evidence of Effects Page (EEP) as a new and effective way to present the evidence for treatment effects. Conclusions: EEPs for treatments of most health conditions can now be developed, and their efficacy in improving informed and shared-decisions can be assessed in suitable trials. O 7.2 Increasing patient comprehension of benefits and risks of medicines using graphics Beyer A1 , Hoekstra T1 , Kingma B1 , Fasolo B2 , Hillege H1 1 University of Groningen, The Netherlands; 2 London School of Economics and Political Science, United Kingdom Background: Graphics are known to be useful in communicating complex information, however they are rarely used in regulatory communication with patients. Objective: To measure patients’ comprehension of benefit and risk of medicines using five presentation formats (text, table, bar graphs, pictograms and survival curves). Methods: Data were collected via a web-questionnaire from patients in the United Kingdom, France and the Netherlands diagnosed with atrial fibrillation (AF), breast cancer (BC) and type II diabetes (DB). Patients were shown presentation formats for benefits and risks of medicines specific to their disease then asked comprehension questions after each presentation. Patients were also asked to indicate their preferred format. Results: We report data for 770 patients (419 DB, 161 AF and 190 BC): age range 46 to 75 years; predominantly male for AF and DB and all female for BC; only 36% to 43% had more than 12 years of education. The table format was most understandable for AF patients with 71% responding correctly to all three questions for both benefits and risks. Among the BC patients the table format was also easier to understand with 82% responding correctly for the risk questions and 70% on the benefit questions. The majority (69%) of the BC patients also responded correctly for the benefit questions when presented in a survival curve. For the DB patients, 80% responded correctly for the risk questionsm but only 54% for the benefit questions when presented with the table format. The results show that across all formats the DB patients did not easily comprehend the benefits. Patients across all disease areas preferred the table format. Among BC patients 49% ranked the table as the most preferred, while for DB and AF it was 43% and 39% respectively. Conclusion: The textual format currently used by regulators to communicate benefits and risks to the public does not result in high comprehension among patients. Regulators should consider adding tables to better support patient comprehension. Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. O 7.3 User-involvement in a Cochrane systematic review update: using structured methods to enhance clinical relevance, usefulness and usability Pollock A1 , Campbell P1 , Baer G2 , Choo PL1 , Morris J3 , Forster A4 1 Glasgow Caledonian University, United Kingdom; 2 Queen MargaretUniversity, UnitedKingdom; 3 UniversityofDundee, United Kingdom; 4 University of Leeds, United Kingdom Background: Active involvement of people with a health condition may enhance the usefulness of systematic review evidence, addressing barriers to uptake of review findings. Challenges to the development of effective methods of involvement within reviews are compounded by poor description and limited evaluation of involvement within many reviews. Objectives: To describe the structured methods used to involve stroke survivors, carers and health professionals in an update of a Cochrane Systematic Review relating to physiotherapy after stroke, and explore the perceived impact of involvement. Methods: After securing relevant funding and ethical approval, we recruited stroke survivors, carers, physiotherapists and educators, and held three pre-planned meetings during a Cochrane Review update. We used formal group consensus methods, based on nominal group techniques, to reach consensus decisions on key issues relating to the structure and methods of the review. Meetings were audio-recorded, and transcribed data used to supplement data on voting decisions. User-group members completed a questionnaire exploring perceived impact of involvement. Results: The user-group comprised 13 people, including stroke survivors, carers and physiotherapists with a range of different professional experience. At Meeting 1 consensus was reached that the methods of categorising interventions used in the original Cochrane Review were no longer appropriate or clinically relevant (11/13 participants disagreed with previous categories). At Meeting 2 group members discussed and reached consensus over a new method of intervention categorisation (12/12 agreed) and at Meeting 3 key messages emerging from the completed review were agreed. All participants strongly agreed that the views of the group impacted beneficially on the review update. They further believed that other Cochrane Reviews would benefit from involvement of a similar user-group. Conclusions: We involved stroke survivors and carers in reaching consensus decisions relating to our review by adopting clearly described, structured methods. We believe that this approach has important implications for future Cochrane Reviews. Cochrane Database Syst Rev Suppl 1–327 (2015) 78 Figure: The conceptual framework of the Ecosystem O 7.4 Trustworthy digital evidence ecosystem to improve diagnosis and treatment of patients Vandvik PO1 , Agoritsas T2 , Kunnamo I3 , Guyatt G2 , Brandt L4 1 Institute for Health and Society, Faculty of Medicine, University of Oslo, Norway; 2 Department of Clinical Epidemiology and Biostatistics, McMaster University, Canada; 3 Department of General Practice and Primary Health, University of Helsinki, Finland; 4 Department of Medicine, Innlandet Hospital Trust-division Gjøvik, Norway Background: Major advances in standards, systems and technological platforms for evidence production and dissemination may together reduce waste and increase value in medical research, reduce information overload and result in better decisions at the point of care. Innovative technological platforms can connect people doing primary research, systematic reviews, guidelines, those creating computerized decision support systems and those involved in quality improvement. Such platforms –which we call an evidence ecosystem –can interact to create, disseminate and implement trustworthy research evidence in clinical practice. Objectives: In this first part of our project to create an evidence ecosystem we developed a conceptual framework and demonstrated its feasibility and relevance through a real life example. Methods: The framework for the Ecosystem uses a PICO (patient, intervention, comparator, outcome)-based shared health data model developed in collaboration with several partners - including Cochrane –that adheres to updated and internationally accepted standards and systems (e.g. GRADE) for trustworthiness. The data-model is implemented in a web-based authoring and publication platform (MAGICapp) used to create, disseminate and dynamically update evidence summaries, recommendations Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. and decision aids. This is integrated with other innovative electronic platforms, e.g. Covidence for key steps in systematic reviews production, and EBMeDS for decision support systems. We applied the ecosystem to address the issue of overtreatment with surgery for meniscal tears. Results: The figure illustrates the ecosystem conceptual framework. We will, at the Colloquium, demonstrate how such an ecosystem can facilitate the processing of high quality evidence from primary research (nationwide observational study and randomized trials of meniscectomies) into systematic reviews, guidelines and decision support tools, followed by quality performance measures and observational studies to document change in practice and outcomes, using overtreatment of meniscal tears as an example. Conclusions: A living evidence ecosystem could improve diagnosis and treatment of patients. Oral session 8 Setting priorities to overcome inequality O 8.1 The global burden of infectious disease as reflected in the Cochrane Database of Systematic Reviews Marzec NS1 , Dmitruk S1 , Boyers L2 , Karimkhani C3 , Czaja CA1 , Trikha R4 , Janoff EN5 , Dellavalle RP6 1 University of Colorado School of Medicine, USA; 2 Georgetown University School of Medicine, USA; 3 Columbia University College of Physicians and Surgeons, USA; 4 Rosalind Franklin University of Medicine and Science, USA; 5 Department of Veterans Affairs Eastern Colorado Healthcare System, USA; 6 Cochrane Skin Group, USA Cochrane Database Syst Rev Suppl 1–327 (2015) 79 Background: As research funding continues to be limited, an approach for setting research priorities is needed. The Cochrane Database of Systematic Reviews (CDSR) publishes rigorous and timely systematic reviews as well as protocols, which together represent the breadth of current and planned prioritization in the Cochrane database. This marker of research priority can be compared to the global burden of disease based on the estimated disease burden measured in disability-adjusted life years (DALYs) reported by the Global Burden of Disease 2010 Project (GBD). This comparison sheds light on current research priorities as well as a possible method of determining future priorities. Objectives: To assess the representation of infectious diseases within the systematic reviews and protocols in the CDSR and compare this to worldwide infectious disease burden as reported by the GBD. Methods: Two investigators independently searched the CDSR database of systematic reviews and protocols for titles, abstracts and keywords related to the infectious Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. diseases reported in the GBD 2010. Reviews and protocols were included or excluded based on whether they added to knowledge regarding the infectious disease that had been searched for. If needed, consensus was reached by a third investigator. The number of reviews and protocols found in the CDSR was used as a surrogate to estimate the existing body of literature regarding each disease and compared to the global burden of the same disease (measured in DALYs). Results: The CDSR was searched for literature relating to 45 infectious diseases based on GBD categorization. These were then divided into three groups: those that were appropriately represented based on DALY (Table 1), those that were under-represented (Table 2), and those that were over-represented (Table 3). Conclusions: Fourteen of the 45 infectious diseases investigated were under-represented in the current literature; 12 were over-represented and 15 were appropriately represented. Infectious diseases research priorities could be informed better using this information. Cochrane Database Syst Rev Suppl 1–327 (2015) 80 O 8.2 Strategies for improving health systems in low-income countries: lessons learnt from four overviews of systematic reviews of health systems interventions Pantoja T1 , Opiyo N2 , Ciaponni A3 , Herrera C4 , Lewin S5 , Oxman A5 , Paulsen E5 , Rada G6 , Wiysonge C7 1 Pontificia Universidad Catolica de Chile, Cochrane EPOC Group, Chile; 2 Child and Newborn Health Group, Kenya Medical Research Institute/Wellcome Trust Research Programme, Kenya; 3 Argentine Cochrane Centre IECS, Institute for Clinical Effectiveness and Health Policy, Argentina; 4 Department of Public Health, Evidence Based Health Care Program, Pontificia Universidad Católica de Chile, Chile; 5 Global Health Unit, Norwegian Knowledge Centre for the Health Services, Norway; 6 Department of Internal Medicine, Evidence-Based Health Care Program, Faculty of Medicine, Pontificia Universidad Católica de Chile, Chile; 7 Centre for Evidence-Based Health Care, Stellenbosch University, South Africa Background: We have conducted four Effective Practice & Organisation of Care (EPOC) overviews of systematic reviews (SRs) of interventions for improving health systems in low-income countries (LICs), including delivery, financial and governance arrangements and implementation strategies. Our objectives were to: (1) provide a broad overview of what is known about the effects of health systems interventions in LICs based on the findings of up-to-date SRs; and (2) identify priorities for new SRs and primary research of such interventions. Objectives: To reflect on the overall findings and lessons learnt from conducting four overviews of systematic reviews of strategies for improving health systems in LICs. Methods: We used structured discussions to identify the key challenges and lessons from conducting the overviews. These discussions took place during face-to-face and online meetings over a 24-month period (March 2013 to March 2015). Results: The four overviews will be published in the Cochrane Library in 2015, and structured summaries of the included reviews will be Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. available at www.supportsummaries.org. The overviews summarize a very substantial body of evidence from Cochrane and non-Cochrane reviews on the effects of interventions for improving health systems. However, there are important differences in the extent of the available evidence across the four overviews, with substantially less evidence available for governance and financial arrangements. Synthesizing the findings of overlapping reviews of the same intervention was challenging. Also, for many comparisons the evidence is of low or very low certainty (mainly due to risk of bias and inconsistency), and for many interventions there are few data on equity outcomes, resource use or unintended consequences. Conclusions: The overviews are an important resource for evidence-informed decision making on health systems in LICs. They are also a resource for identifying important gaps in the body of evidence for this field and have informed priorities for new reviews and research. O 8.3 Using priority setting to enhance the policy relevance of systematic reviews: a case study from the South African Initiative for Systematic Reviews on Health Policies and Systems Odendaal W1 , Daniels K1 , Lewin S2 , Kredo T3 , Young T4 1 South African Medical Research Council, South Africa; 2 South African Medical Research Council; Norwegian Knowledge Centre for the Health Services, South Africa; Norway; 3 South African Cochrane Centre, South African Medical Research Council, South Africa; 4 Centre for Evidence-Based Health Care, University of Stellenbosch; South African Cochrane Centre, South African Medical Research Council, South Africa Background: Identifying and prioritising systematic review topics that address stakeholders’ interests is not easy, particularly when time and resources are constrained. The Alliance for Health Policy and Systems Research recently Cochrane Database Syst Rev Suppl 1–327 (2015) 81 funded a two-year initiative in South Africa to conduct four systematic reviews on health systems questions. Central to the initiative was engagement with stakeholders to identify and select priority review questions. Objectives: To share lessons learned in engaging stakeholders and identifying their priorities for new systematic reviews. Results: We used a multi-step process to identify priority topics for reviews. We invited 154 public health policymakers, managers, providers and researchers across sub-Saharan Africa to submit priority topics for health system reviews. Ninety topics were received. A core team then grouped these into broad themes; transformed these themes into systematic review questions; and searched for existing reviews on these questions. This resulted in a shortlist of 39 review questions. A small advisory panel then used an iterative process to select four review topics. This included ranking each question as high or low priority and discussions within the panel to resolve differences in ratings. Conclusions: The strengths of this prioritisation approach included: (1) rapid engagement of a wide range of health stakeholders; and (2) the identification of review questions that address important national and regional health systems questions. A key challenge was transforming the themes into review questions as, (1) many of the suggested topics were not questions that could best be addressed using systematic review methods; (2) some topics encompassed multiple questions that had to be split into reviewable questions; and (3) several topics were partly covered by existing reviews. In addition, the final shortlist of review questions was not verified with those who proposed these topics, and the questions may therefore have diverged from the original topic. O 8.4 Building a systematic review movement in the humanitarian field Ott EM1 Oxfam GB, United Kingdom 1 Background and objectives: Humanitarian actors and researchers have amassed evidence about the state of knowledge in the humanitarian sector, including grey literature assessments of what works and what does not. Synthesizing this information and gaining research uptake by policymakers and humanitarian practitioners remains challenging. The Humanitarian Evidence Programme commissions a series of systematic reviews to distil evidence in the humanitarian sector and focuses on uptake and communication of findings to key stakeholders, with the ultimate goal of improving humanitarian policy and practice. The programme is a partnership between Oxfam GB and Feinstein International Center (FIC) at Tufts University, funded by UK Aid. This partnership is unusual in its emphasis on practitioners and the symbiotic involvement of individuals occupying multiple roles in the humanitarian field. Methods: The programme collected data for deciding the review questions and research uptake strategy through a survey with 273 responses from individuals based in 55 different Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. countries with an average of 15 years of experience as well as 45 key informant interviews. Results: The data show that key humanitarian practitioners access and use research through conversations with key colleagues as well as through briefings, articles, and the media. Importantly, some individuals act as ‘nodes’ through which a number of individuals access trusted research. The resulting research uptake strategy for the programme uses three core approaches: 1. meaningful consultation throughout the life process of the programme; 2. targeting of key thought leaders; and 3. leveraging existing platforms and using a multi-channel communications strategy to increase reach. Conclusions: For the Humanitarian Evidence Programme, this research contributed to an iterative process to develop a list of systematic review questions in the humanitarian field and a research uptake plan. The first call for proposals solicited a strong response, showing interest in this programme. For the broader systematic review field, this model may be of interest and adaptable for other programmes and reviews. Oral session 9 Systematic review and meta-analysis methods O 9.1 Extent and handling of missing dichotomous outcome data in 100 Cochrane and non-Cochrane systematic reviews Kahale LA1 , Diab B1 , Brignardello-Petersen R2 , Mustafa R3 , Busse JW2 , Agarwal A2 , Kwong J4 , Li L4 , Neumann I5 , Lopes LC6 , Olav Vandvik P7 , Briel M8 , Iorio A2 , Guyatt GH2 , Akl EA1 1 American University of Beirut, Lebanon; 2 McMaster University, Canada; 3 University of Missouri-Kansas City, USA; 4 Sichuan University, China; 5 Pontificia Universidad Católica de Chile, Chile; 6 University of Sorocaba, Brazil; 7 Norwegian Knowledge Centre for the Health Services, Norway; 8 University Hospital of Basel, Switzerland Background: Reports of randomized clinical trials (RCTs) do not typically report whether certain categories of participants (e.g. non-compliers, those who withdraw consent) were followed-up for the outcome(s) of interest. It is not clear how systematic reviews (SRs) address this lack of information, i.e. assume they have missing participant data (MPD) or not. Objectives: To describe how SRs report and handle categories of participants that could be potentially counted as having MPD. Methods: We included SRs reporting a group-level meta-analysis of a patient-important dichotomous efficacy outcome, with a statistically significant effect estimate. Thirteen reviewers, working in pairs, independently extracted data from eligible SRs. We focused on 10 categories that could be counted as potentially having MPD: ‘ineligible participants’, ‘did not receive any treatment’, ‘withdrew consent’, ‘outcome not assessable’, ‘explained and unexplained lost to follow-up’, ‘dead’, ‘experienced adverse events’, ‘non-compliant’, ‘discontinued prematurely’, and Cochrane Database Syst Rev Suppl 1–327 (2015) 82 ‘cross-over’. We assessed whether the systematic reviewers: (1) planned in the methods section to collect information about those categories; (2) reported in the results section, the number of participants in those categories; and (3) reported on methods for handling MPD in their primary meta-analysis. Results: We included a random sample of 100 Cochrane and non-Cochrane SRs published in 2012. Tables 1–3 report our main findings. Twenty-five per cent of the SRs reported a plan to collect information about the categories of interest in their methods section, whereas 45% reported in their results section the number of participants in those categories. Only 13% reported a method to handle MPD in their primary meta-analyses, which were primarily complete case analysis. Conclusions: Most SRs do not explicitly report sufficient information on categories of participants that could be potentially counted as having MPD, or handle MPD in their primary meta-analysis. O 9.2 The changing world of data sharing and data transparency: what does this mean for individual participant data reviews? Nolan S1 , Tudur Smith C1 , Marson A1 1 University of Liverpool, United Kingdom Setting: There are many advantages to undertaking an individual participant data (IPD) meta-analysis compared to a summary data approach. IPD meta-analysis allows more flexible, complex statistical exploration of data, often increasing the number of clinical questions that can be addressed. In many contexts where summary data are insufficient or unavailable, an IPD analysis is the only feasible approach. However, procedures for obtaining IPD are time-consuming, resource intensive and often unsuccessful due to unavailability of data or concerns regarding patient privacy, collaboration and data sharing. Many initiatives promoting data sharing and data transparency within the pharmaceutical industry have developed quickly in recent years; such changes could have a large impact on IPD meta-analysis. Background: From 1997 to present, the Cochrane Epilepsy Group has undertaken IPD meta-analyses and an IPD network meta-analysis is currently being updated. For all studies identified as eligible for these reviews, an IPD request was made to the lead or corresponding author or the sponsoring organisation if a study was industry-funded. Methods of contacting relevant data providers, requirements of an IPD request, content of data provided and success rates of IPD requests have changed dramatically over time. The launch of ClinicalStudyDataRequest.com (CSDR) platform and the SAS multi-sponsor data access environment in the last two years to allow researchers to request and remotely analyse de-identified IPD has been a significant change. Discussion: IPD requests from 40 studies eligible for the IPD network meta-analysis from 2011 to the present will be discussed including requests pre-, during, and post the launch of CSDR. Advantages and practical issues of CSDR and the SAS data access system will be presented and the potential impact of these new initiatives on future IPD analyses will be discussed. Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Cochrane Database Syst Rev Suppl 1–327 (2015) 83 O 9.3 Multivariate meta-analysis of multiple correlated outcomes with individual participant data: how much do we gain? Frosi G1 , Riley R2 , Williamson P1 , Kirkham J1 1 Biostatistic Department of University of Liverpool, United Kingdom; 2 Research Institute of Primary Care & Health Sciences, Keele University, United Kingdom Background: Missing treatment effect estimates for particular outcomes in a study have the potential to affect the conclusions in a meta-analysis, especially if missingness is a result of outcome reporting bias (ORB). As well as missing treatment effect estimates at the study level, outcome data may also be missing within studies at the individual participant level. Multivariate meta-analysis of individual participant data (IPD) has the potential to overcome the impact of both these problems, by utilising the correlation between outcomes. Objectives: To investigate, in a range of ORB and missing data scenarios, the magnitude of bias in pooled treatment effect estimates for multiple outcomes using standard (univariate) meta-analysis, and to quantify how much the ‘borrowing of strength’ (BoS) from multivariate meta-analysis reduces such bias and increases precision. Methods: A simulation study was conducted where IPD was generated from an assumed multivariate fixed-effect model, and missing data were created at either the patient-level or the study-level, or both. In each simulation, the bias, precision and coverage of univariate and multivariate methods was compared, and the BoS quantified. Results: Results show that the BoS in a multivariate model can substantially reduce the magnitude of bias and increase precision in the pooled estimates, especially when ORB is present and when correlation is modelled at both the patient-level and the study-level. In a missing at random scenario (0.8 w/s correlation) the BoS was 34.9% (outcome 2), meaning a reduction of 35% in the variance of the pooled estimate. Moreover BoS increases as the correlation and amount of missing data increases. Conclusions: Meta-analysis results may be unreliable if there are missing outcome data. A multivariate meta-analysis approach is a potential statistical solution for reducing the impact of missing data, and is especially appealing when IPD are available to deal with missing data at both the patient and study levels. O 9.4 Evaluation of the consistency between network meta-analysis and standard pairwise meta-analysis: a meta-epidemiological study Yuan J1 , Fu X1 , Mao C1 , Yang Z1 , Tang J1 , Yuan J1 , Zhang Y1 1 Division of Epidemiology, School of Public Health and Primary Care, The Chinese University of Hong Kong, Hong Kong, China Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Background: Network meta-analysis is a relatively new type of data synthesis statistics, there is still considerable concern about its validity. Objectives: To evaluate the consistency between standard pairwise meta-analysis and network meta-analysis and explore the potential factors that are associated with the inconsistency. Methods: PubMed, the Cochrane Library, EMBASE were searched and reference lists of relevant methodological reviews were checked to identify network meta-analyses that reported the estimate effects of both standard pairwise meta-analysis and network meta-analysis, or provided raw data to allow us to calculate direct estimates. We assessed both quantitative consistency (whether the difference between the matched estimates is statistically significant) and and qualitative consistency (whether the direction or statistical significance of the paired estimates is different). The association between inconsistency and potential influential factor was tested by the χ 2 test. Results: A total of 1901 matched estimates of standard pairwise meta-analyses and network meta-analyses were obtained from 90 studies. Of which 20 (1.1%, 95% confidence interval (CI) 0.6% to 1.5%) and 367 (19.3%, 95% CI 17.55% to 21.1%) matched estimates showed quantitative inconsistency and qualitative inconsistency, respectively. The qualitative inconsistency was significantly associated with continuous data (P value 0.002), random-effects network meta-analysis model (P value 0.045), closed loop in the network (P value 0.001), invalid coherence assumption (P value 0.036), and high risk of publication bias (P value 0.019). The quantitative and qualitative inconsistency rates for matched estimates of indirect comparisons and standard pairwise meta-analysis were 9.1% (5/55) and 32.7% (18/55), respectively. Conclusions: Network meta-analysis is a reliable statistical procedure if it is done correctly. It hardly shows any quantitative inconsistency when compared to standard pairwise meta-analysis. However, qualitative inconsistencies exist in about one-fifth of the estimates, which should be interpreted with caution. Rapid oral session 13 Translations RO 13.1 One year in: facts and figures about the impact of Cochrane’s translation strategy Ried J1 , Hassan H2 1 Cochrane Central Executive, Translations Co-ordinator, Germany; 2 Cochrane Central Executive, Translations Support Officer, United Kingdom Background: Cochrane approved a translation strategy in 2014 with the aim of increasing the capacity and providing central coordination for sustainable Cochrane translation activities in a broad range of languages. The translation strategy contributes to strategic goals 2 and 3: ‘Making Cochrane Database Syst Rev Suppl 1–327 (2015) 84 our evidence accessible’; and ‘Advocating for evidence’. Objectives: To describe the main components of Cochrane’s translation strategy, and to show what impact it has made to date on translation and related activities. Methods: The analysis will consist of the collection of metrics as well as a brief quantitative and qualitative survey with our language project managers, and will investigate the following aspects: - the number of languages we are translating into, and their importance on the global linguistic landscape; - how much and what content has been translated; - how many people are involved in translations, and in what capacity; - how much and what type of resources the projects have; - how productive the projects are; - how sustainable the projects are; - how translations may have led to local outreach and dissemination activities; - how translations may have led to a local increase of interest in Cochrane (e.g. access to our content, media coverage); - whether translations may have led to new funding or partnerships. Retrospective comparison data will be provided where possible. Results: The results from the analysis will be available for the Colloquium. As of March 2015, we have published more than 12,000 translations of Cochrane summaries across 12 different languages. The majority of our translation teams rely on low resources and volunteers. Most new projects require several months to prepare and then reach a certain continuity, and a number of teams have successfully reached this point even without any dedicated funding. Conclusions: While the implementation of the translation strategy is still at its early stages, increase in production, languages and access already suggests that our approach has a lot of potential to increase our capacities even in low resource settings, and with relatively low central investment. members of Cochrane Croatia. Crowdsourcing was identified as a potential low-resource approach to increase translation capacity. Objectives: To test the involvement of volunteer medical students and health professionals in translating PLS into Croatian and to measure the impact on production of translations. Methods: In July 2014, we started inviting students and university colleagues to participate as volunteers in our translation project. A public call was also published through the Cochrane Croatia Facebook page, Cochrane Croatia website and in the University newspapers. Medical school teachers incorporated translation of one PLS as a mandatory part of the exam in research methodology courses, and some of the students volunteered to translate more after the exam. We developed supporting material describing the translation process and guidelines. All translations submitted by the volunteers are checked and edited by Cochrane Croatia’s volunteer staff to ensure accuracy and consistency. Volunteers are acknowledged along with the published translations. Results: Fifty-six volunteers signed up to contribute translations (as of February 2015). In the first 18 months of the translation project, prior to adopting a crowdsourcing approach, we translated about 200 PLS. Within the first six months of crowdsourcing we were able to complete 350 additional PLS translations–about five times more than before. Volunteer time of the two project administrators devoted to the translation project approximately doubled after introduction of crowdsourcing. Conclusion: Involving volunteers using a crowdsourcing approach and assigning translations to students as a part of their medical curriculum substantially increased our capacity to translate Cochrane evidence, but it required significantly more volunteer time from the translation managers. RO 13.2 RO 13.3 Involving volunteers in translating Cochrane Plain Language Summaries to increase capacity The importance of critical revision of translated abstracts Puljak L1 , Ried J2 1 Cochrane Croatia, Croatia; 2 Translations Co-ordinator, Cochrane Central Executive, Germany Background: Cochrane Croatia started translating the Plain Language Summaries (PLS) of Cochrane Reviews into Croatian in 2013 to overcome the language barrier. Initially, translations were mainly completed by two volunteer staff Logullo P1 , Albuquerque JV1 , Porfirio G1 , Martimbianco ALC1 , Riera R1 , Atallah AN1 , Torloni MR1 1 Brazilian Cochrane Center, Brazil Reviewing the translations of abstracts done by volunteers: is it worth the effort? Experience from the Brazilian Cochrane Center: Background: Since 2013, volunteers from the Brazilian Cochrane Center (BCC) have translated English Cochrane Abstracts into Portuguese; these are uploaded to Table 1 Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Cochrane Database Syst Rev Suppl 1–327 (2015) 85 the Cochrane Library only after being reviewed by a language and a content expert. English belongs to the Germanic family of languages, while Portuguese is part of the Italic language family, therefore, the construction of sentences is different, as well as the use of double negatives, the choice for verbal tenses and the order of words. These differences make translating Cochrane Abstracts a challenging task. To help translators, we created and distributed a short guide on the most common mistakes in translating these abstracts to Portuguese. Objectives: To assess the most frequent Portuguese language errors found in abstracts translated by BCC volunteers. Methods: We analyzed the last 12 abstracts translated by volunteer physicians, psychologists, physical therapists and a dietitian. A Portuguese language expert experienced in reviewing medical texts and a physician reviewed all abstracts for language and content accuracy. Language errors and corrections were classified by type and frequency. Results: Most corrections (mean 15.75/translation) concerned style and were made to improve text readability. However, all 12 texts also had grammatical (12.37/translation) as well as misspelling errors or ’typos’ (11.85/translation; Table 1). Conclusions: Although the volunteer translators are highly motivated and qualified health professionals, their translations can be considerably improved by a professional language review. Since translated Cochrane Abstracts are read by thousands of professionals and consumers worldwide, it is important to use a review process to ensure not only that their content is accurate, but also that the text is grammatically correct, and is easy to read and understand for all users. This language review is important and should be encouraged in other languages to ensure the quality of translations produced by Cochrane. for this project. Objectives: To create usable, reliable, fully automatic translation of public health information, initially testing with translation from English into Czech, Polish, Romanian and German. Methods: We will use recent advances in machine translation (MT) to create a system for the automatic translation of public health information, with a focus on preservation of meaning. We will include recent work on domain adaptation, translation into morphologically rich languages, terminology management, and semantically enhanced MT to build a reliable system for the health domain. We will iterate cycles of incorporating improvements into the MT systems annually, with careful evaluation and user acceptance testing. We will develop metrics to evaluate the quality and measure the impact on post-editing of the obtained results in each cycle. Cochrane and NHS24 content will also be translated in each cycle and published on their websites, and users invited to participate in a survey to give feedback on the quality and usability of the translations. Web usage statistics will be analysed. Results and Conclusions: The first version of the MT system is to be deployed in Sept 2015. Preliminary results and conclusions will be presented at the Colloquium as available. The focus will be on process and the Cochrane use case, rather than technological details. RO 13.5 Lessons from translating Cochrane Reviews in nursing into Korean Park M1 , Lee M1 , Jang I2 1 College of Nursing Chungnam National University, South Korea; 2 Chungnam National University Hospital, South Korea RO 13.4 Health in my Language: health domain adapted machine translation as a means to tackle the resource issues for translation Ried J1 , For the HimL consortium N2 1 Cochrane Central Executive, Translations Co-ordinator, Germany; 2 University of Edinburgh, Charles University Prague, Ludwig Maximilian University of Munich, NHS 24, Cochrane, and Lingea, NA Background: To an ever-increasing extent, web-based services are providing a frontline for healthcare information in Europe. They help citizens find answers to their questions, and they help them understand and find the available local services. However, due to the number of languages spoken in Europe, and the mobility of its population, there is a high demand for these services to be available in many languages. In order to satisfy this demand, we need to rely on automatic translation, as it is infeasible and unaffordable to translate manually into all languages requested. Health in my Language (HimL) is an EU-funded, three-year project, aiming at addressing this need. Health information produced by Cochrane and NHS24 (health information and self care advice for the people of Scotland) will serve as the test case Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Background: There are current projects to translate Cochrane content into various languages to improve the availability of its content. In Korea, we started collaborated efforts (medicine, oriental medicine, nursing, and public health area) to translate Cochrane Reviews using professional translation and editing. Objectives: This paper describes the experiences of translating Cochrane Reviews in the nursing area into Korean and discusses lessons arising from the experience. Translation process: We set the high priority topics in the nursing arena which needed to be translated first. In nursing, the priority was given to recently published reviews in gerontology and evidence-based nursing. These translations were complemented by the Smartling system, which makes working together easier for the project manager and the translator. Lessons learned: It is important that the objective of the translation (literal translation, conceptual translation, or culturally equivalent translation) be explicitly identified before the translators begin their work. While the translated reviews are grammatically correct and follow the wording of the original, it is hard to translate the professional language into plain Korean for lay people. There is a need for an agreement of standardized language in each specialty area. An interdisciplinary standardized language system needs to be used. The translation work is time- and effort-consuming, Cochrane Database Syst Rev Suppl 1–327 (2015) 86 it is highly recommended that the priority list of reviews is set up with the members’ agreement. It would be helpful to make a system that facilitates review of the translation by peers and lay people. A high percentage of the reviews in nursing concluded that the effects of certain interventions were not clear. It would be a good strategy to prioritise those reviews with a clear conclusion. Acknowledgment: This work was supported by Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Education, Science and Technology (No. 2010-0024922). RO 13.6 Russian translations of Cochrane Plain Language Summaries: experience with and feedback from an online survey on quality and importance Ziganshina LE1 , Yudina EV1 1 Kazan Federal University, Affiliated Centre in Tatarstan of the Nordic Cochrane Centre, Russian Federation Background: In 2014 Cochrane initiated its translation strategy. The Russian translation project was started in May 2014 by a team of volunteers from the Kazan Federal University, an affiliated center in Tatarstan of the Nordic Cochrane Centre. Objectives: To assess the quality of Russian translations of Cochrane Plain Language Summaries (PLS) and their potential impact on the Russian speaking community through user feedback with the aim of furthering the translation project. Methods: We conducted an online survey (3-24 March 2015) via Google Docs. We invited respondents through the electronic Russian language discussion forum on Essential Medicines (E-lek), links to a survey on the Russian cochrane.org website, and invitations to Cochrane contributors registered in Archie from potential Russian-speaking countries. We set up the survey in Russian and English. Results: By 24 March 65 people had taken part in our survey (56 in Russian, 9 in English), mostly representing health professions (n = 48; 74%). The Russian text of translations and basic meaning were clear to the majority of respondents (n = 63; 97%) to varying degrees. The respondents rated the quality of translations as ’excellent’ (n = 20; 31%), ’good’ (n = 33; 51%) and ’satisfactory’ (n = 9; 14%). All respondents noted good compliance of the Russian translation with the original English text. Nearly all of them recognized the need for Cochrane evidence for Russia and Russian-speaking countries (n = 63, 97%), for their work/school/life (n = 61; 94%) and the potential impact on their practice/attitude to drugs or diagnostic procedures (n = 60; 92%). Many respondents (n = 42; 65%) preferred translation texts worded in Russian without precise compliance to the original text; 24 respondents volunteered to become members of the translation project. We received valuable suggestions for improvement and further development of the Russian translation project. Conclusions: The survey provided positive feedback so far. We will continue Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. the survey to confirm or adjust our findings as we translate more PLS and to reach out to a wider audience and to attract new volunteers to work in the project. RO 13.7 Strategies for eliciting mutual benefit from translating Cochrane Review content: the example of Cochrane Kompakt in German Toews I1 , Voigt-Radloff S1 , Flatz A2 , Nussbaumer B3 , Meerpohl JJ1 , von Elm E2 1 2 Cochrane Germany, Germany; Cochrane Switzerland, Switzerland; 3 Cochrane Austria, Austria Background: Cochrane Reviews provide health information in an era of information overload. With its ’Strategy to 2020’, Cochrane has approved a translation strategy and established an infrastructure to support translation projects. The German-language translation project Cochrane Kompakt (www.cochrane.org/de/kompakt) is a collaboration between Cochrane entities in Switzerland, Austria and Germany and actively reaches out to the health professions and their organisations to maximise benefit from the limited resources available. Methods: We work with health professionals, who are interested in disseminating Cochrane Reviews, to translate Cochrane Plain language summaries into German on an ongoing basis. Through their networks, they have access to domain-specific audiences and dissemination channels. The cooperation we target consists of them lending us their expertise for translation; in return translated summaries are co-published in German-language discipline-specific journals. This increases the dissemination and impact of translated Cochrane summaries. Via workshops and seminars, opportunities to get involved in Cochrane Kompakt are promoted. A standard approach to instruct translators was developed to ensure quality and consistency. The coordinators introduced new translators to the web-based translation platform, Smartling, by using tailored instructions, and provided them with ongoing assistance. All texts are redacted by at least two project collaborators. Results: As of March 2015, we published 174 Cochrane Review summaries in German on Cochrane Kompakt. Of those, 25 were co-published in scientific journals (e.g. pt Zeitschrift für Physiotherapeuten), and another 12 are scheduled for publication in 2015. Mutual benefit for both Cochrane and the involved health professions increases the number of translation resources and translations, keeps up motivation and has a snowball effect with current translators promoting their own work and recruiting additional volunteers. It is worthwhile for other translation teams to consider this strategy. Active involvement of health professions will ensure that translated content is disseminated to target readerships. Cochrane Database Syst Rev Suppl 1–327 (2015) 87 RO 13.8 Enhancing dissemination of Cochrane evidence via German translations of physiotherapy-related Cochrane Plain Language Summaries - a collaborative project Braun C1 , Bossmann T2 , Ehrenbrusthoff K3 , Lohkamp M4 , Jahnke N5 , Handoll H6 1 Hochschule 21, Department of Health (Physiotherapy), Buxtehude, Germany; 2 Technische Universität München, Department of Sport and Health Sciences, Germany; 3 Hochschule für Gesundheit, Department of Applied Health Sciences, Bochum, Germany; 4 SRH Hochschule, School of Therapeutic Sciences, Heidelberg, Germany; 5 Cochrane Cystic Fibrosis and Genetic Disorders Group, Liverpool, United Kingdom; 6 Cochrane Bone Joint and Muscle Trauma Group, Manchester, United Kingdom Background: The exponential increase in health information constitutes a universal challenge, more so for non-English speakers as new scientific knowledge is almost exclusively published in English. As part of its Strategy to 2020, Cochrane is working to ‘‘make Cochrane evidence accessible and useful to everybody, everywhere in the world’’. A key focus has been to translate Cochrane abstracts and plain language summaries (PLS) into different languages, which poses various challenges as a resource-intensive task requiring both linguistic and context expertise. Objectives:- To provide German translations of the physiotherapy-related PLS of 10 Cochrane Reviews by the Bone, Joint and Muscle Trauma Group (BJMTG). - To examine and document the translation process and provide feedback to BJMTG on aspects related to the translation and content of the PLSs. Methods: We are building on our ongoing initiative to provide German translations of PLS of physiotherapy-related Cochrane Reviews on Cochrane websites (www.cochranelibrary.com and www.cochrane.org/de) and in the German physiotherapy journal Zeitschrift für Physiotherapeuten. A team of German physiotherapists, all experienced in translating PLS, with input from a native English advisor, translate the PLS of the selected BJMTG Reviews. Translation follows a standardised process involving at least two translators per PLS and the use of the translation software Smartling R. Feedback is documented on a pre-developed form and includes aspects such as difficulties experienced when translating the PLS into German plain language and the completeness of the PLS. The translation work is funded by BJMTG. Preliminary results and prospects: The project is ongoing; it started in February 2015. We will report our findings at the Colloquium. So far (March 2015), two translations have been completed. A third one is underway. We hope that insights gained from our project Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. will stimulate and facilitate the involvement of professional groups in other Cochrane translation projects, particularly for languages outside larger-scale Cochrane translation projects. RO 13.9 Using social media to disseminate translated Cochrane evidence and to promote high-quality health information in non-English speaking regions Puljak L1 1 Cochrane Croatia, Croatia Background: Social media have emerged as a potentially useful tool for knowledge transfer (KT) in recent years. To date, however, there is little information about the impact of dissemination of translated Cochrane Plain language summaries (PLS) via social media tailored to regional audiences. Objectives: To test and evaluate the use of a Croatian Facebook page and Twitter account as dissemination tools for Croatian PLS translations. Methods: In early 2013, Cochrane Croatia started translating PLS into Croatian to overcome the language barrier. In March 2013, a ‘Cochrane Croatia’ Facebook page was set up. It was mainly used to post new PLS translations and other relevant information in Croatian. A Croatian ‘Cochrane Health’ Twitter account was set up as well in April 2014 for the same purpose. Results: To date, the Cochrane Croatia Facebook page has attracted 1858 followers (February 2015), mostly from Croatia and neighboring countries with similar languages. Most of the followers are in the 25-44 years age group. The most popular PLS were related to pregnancy, childbirth, dental care, occupational health, depression and prostate cancer. The Facebook page has enabled direct interaction with our audience, which allowed us to determine that our followers include lay persons, health professionals and journalists. As a result of the Facebook activities, we experienced a substantial increase in media uptake of Cochrane stories from an average of one per month before Facebook activities to about 20 per month now (February 2015). The popularity of Twitter in Croatia is not comparable to that of Facebook. As a result, the Twitter account attracted only 34 followers. Conclusion: Facebook proved successful in creating multiple opportunities to engage with our Croatian target audience, to disseminate evidence-based health information, and to increase the uptake of Cochrane evidence by regional media. Cochrane groups should however focus on the most popular social media tools in their setting for KT purposes. Cochrane Database Syst Rev Suppl 1–327 (2015) 88 Rapid oral session 14 Education and training RO 14.2 Train the trainers: Medical Library Literacy for Health Professionals in Ethiopia RO 14.1 Building capacity for systematic reviews in low-income countries beyond the classroom: the Africa centre for systematic reviews and knowledge translation Obuku E1 , Kinengyere A1 , Ssenono R1 , Sewankambo N1 1 Africa Centre for Systematic Reviews and Knowledge Translation (Africa Centre), Makerere University College of Health Sciences, Uganda Background: Researchers in sub-Saharan Africa have limited knowledge and skills to conduct systematic evidence syntheses. The ’Africa Centre’ at Makerere University, College of Health Sciences Uganda was set up to address this gap in the East African region. Objectives: To describe our experience in building teams of systematic reviewers in Uganda and the East African Community. Methods: Design: We administered a sequential three-phase course to two groups of scientists from 13 research and teaching institutions in Africa (started Aug 2013 and Dec 2015). These modules consisted of a self-driven introductory learning about systematic reviews for 10 and 21 days respectively: phase 1: face-to-face didactic lectures and hands-on group work for five days at Makerere University, Kampala, Uganda; phase 2: technical support (Skype calls, online metorship) in conducting systematic reviews until publication of the review report (>18 months). Outcome measures: Systematic reviewing activities including: question identification, protocol development, protocol registration, review execution, publication, and grant application. Results: Participants (n = 51) came from eight countries including Rwanda (4), Ethiopia (1), Botswana (3), Southern Sudan (1), Tanzania (4), Kenya (4), Cameroon (1) and Uganda (33). Twenty-one (41%) were female scientists. Forty-eight (94%) and 51 (100%) completed the self-driven introductory learning and face-to-face sessions respectively. During this period, participants identified 16 new potential review questions. Three new protocols have been registered in PROSPERO and one with Cochrane HIV group; three new protocols are under development whilst three older protocols were revived and under conduct. Only one review is published and only one of three new systematic review grant application attempts have been successful. Conclusions: These preliminary results depict some progress in the quest to build systematic review teams in the East African Community. In our next steps we aim to collaborate with the South African Cochrane Centre to conduct a data synthesis (GRADE: Grading of Recommendations Assessment) and manuscript writing workshop. Kendall S1 , Hagstrom C2 1 Mount Sinai Hospital, Canada; 2 University of Toronto, Canada Background: In response to the Ethiopian Government’s aim to train 5000 specialist MDs and PhDs and 10,000 Masters graduates by 2018, the Toronto Addis Ababa Academic Collaboration (TAAAC) was established to co-build capacity and sustainability in graduate programming at Addis Ababa University (AAU).Enhancing the discipline of medical library sciences is essential to the successful support of this expansion. Objectives/methods: With international collaboration, a program for clinical medical librarians will build capacity in library literacy skills, including the accessing of up-to-date information, with foundational knowledge of appraising the evolving literature in evidence-based medicine and critical thinking skills to support medical faculty and trainees. TAAAC-Library Sciences program has embarked on a plan for a prolonged partnership between the University of Toronto (UofT) and AAU by supporting library services with annual train-the-trainer programs, and the provision of access by affiliated researchers to the UofT libraries (Ptolemy) to access online resources including clinical decision support tools. Results: Responding to the needs outlined by AAU, librarians and library assistants participate in workshops to upgrade clinical library skills and then pass this knowledge to others. We have begun to assess the impact of our library training sessions and the outcomes of this transfer of knowledge. We do know that librarians from UofT play a valuable role in supporting and assisting the learning of Ethiopian librarians and in return benefit from a wider understanding and experience of knowledge translation skills. Conclusions: We have now taught over 700 learners. These learners include library staff, physicians, medical and nursing students. Our next steps include the beginnings of our distance or remote library search support, a review of the use of the UofT Libraries’ e-resources, assignment based scenarios for clinicians, and web-based instructional modules to complement in country training sessions. RO 14.3 Spread of evidence-based practice into directors of nursing by an outreach campaign in Taiwan Weng Y1 , Kuo K2 , Chiu Y2 , Chen K2 , Chen C2 1 Chang Gung Memorial Hospital, Taiwan; University, Taiwan 2 Taipei Medical Background: Directors of nursing (DONs) have important influence in the dissemination of evidence-based practice (EBP) in the hospital settings. Nevertheless, diffusion of EBP into DONs was not well surveyed. Objectives: A nationwide program provided information resources and promotional Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Cochrane Database Syst Rev Suppl 1–327 (2015) 89 activities of EBP in regional hospitals of Taiwan since 2007. The current study aimed to evaluate the impacts of this campaign on DONs. Methods: A cross-sectional questionnaire survey for DONs was conducted in 2007, 2009 and 2011 to examine views related to EBP, including changes in beliefs, attitudes, knowledge, skills, and barriers. Chi-square test was used for statistical analyses (SPSS 19.0 for Windows, SPSS, Chicago, IL, USA). Results: This study enrolled 267 DONs in 2007, 257 DONs in 2009, and 287 DONs in 2011. During the four-year study period, their knowledge and skills of EBP increased (P value < 0.001). However, their beliefs and attitudes were not significantly changed. A majority of barriers significantly declined, including no capable designated personnel (P value < 0.001), deficient convenient application kits (such as personal digital assistants and brochures; P value < 0.001), insufficient time (P value < 0.001), inadequate basic knowledge of EBP (P value < 0.001), lack of skill in literature searching (P value = 0.043), and insufficient library resources (P value < 0.001). Conclusions: A multifaceted campaign can increase the knowledge and skills of DONs and reduce their barriers toward EBP. Spreading EBP simultaneously with a mass media campaign may serve as a key element to accelerate its implementation. The data suggest an outreach intervention is useful to disseminate the implementation of EBP for DONs. RO 14.4 Students 4 Best Evidence: ‘What’s the evidence for this?’ campaign Millward H1 1 Cochrane UK, United Kingdom Introduction: Students 4 Best Evidence (S4BE) is an online community for students interested in evidence-based health care; it aims to help students learn more about evidence-based practice and the methodological concepts underpinning it. S4BE involves students from school age to university through relevant, useful resources as well as being a space for them to communicate their knowledge and interact with fellow students. Aims: In April 2015, S4BE is launching a year-long campaign to ask, ‘What’s the evidence for this?’ The aim of the campaign is to ignite student champions globally, to fight for evidence, to spread understanding on the use of evidence and show how it can improve global health. We will do this through a series of mini events and projects throughout the year. We would like to invite Colloquium attendees to join the campaign and share the ‘S4BE What’s the evidence for this?’ campaign with their student networks. Results: We want the campaign to result in improved understanding of evidence and the concepts used in evidence-based practice. We seek to strengthen the S4BE community with a new group of student champions, willing to ask ‘What’s the evidence for this?’ Conclusions: We would like to build on the success of blogs like, ‘A beginner’s guide to interpreting odds ratios, confidence intervals and P values’, that has been viewed over Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. 100,000 times since publication in August 2013 and make sure students know how to campaign successfully for best evidence. RO 14.5 ‘Cochrane Lehrmodule’: development and implementation of teaching modules in university-affiliated courses Mahlknecht P1 1 Cochrane Austria, Austria Background: The Danube University Krems, where Cochrane Austria is hosted, is offering a variety of courses in the field of medicine and health sciences. Within these courses, modules on evidence-based research skills and methods have gained popularity. Cochrane Austria was requested to develop teaching modules on evidence-based methods and contents to meet their increasing demand. This is in line with the Cochrane Strategy to 2020 and the Cochrane training and professional development strategy, including the establishment of university-affiliated courses, and providing appropriate training opportunities. Objectives: To develop teaching modules and implement them in university-affiliated courses, transferring research skills and basic knowledge in the area of evidence-based healthcare. All modules follow a standard structure that can be tailored to the audience of the various courses. The modules are also intended for people without prior knowledge of evidence-based research methods. Development and structure: In consultation with the heads of departments and course directors, needs, requirements, and previous knowledge of the students are assessed in order to turn expectations into realistic and effective learning goals. A challenge is to define the level of knowledge and skills that is essential (compulsory modules) or optional (elective modules) for the general audience to achieve. Depending on learning goals, compulsory modules are delivered in blocks of five days and include evidence-based medicine and healthcare, facts about Cochrane, epidemiology, statistics, study designs, critical appraisal of studies, including qualitative research and guidelines, and literature searches. Elective modules are implemented based on the background, interests and needs of the respective audience. Prospects: The first block of teaching modules will be launched end of April. The purpose of this presentation is to highlight issues around the implementation of the training, share teaching experiences so far, and discuss potential modifications to the modules. It will be of interest to a wide audience involved in training and teaching evidence-based research methods. Cochrane Database Syst Rev Suppl 1–327 (2015) 90 RO 14.6 Online systematic review methods training resources: environmental scan and identification of key characteristics Parker RM1 , Visintini S2 , Ritchie KC3 , Hayden JA1 1 Dalhousie University, Canada; 2 Maritimes SPOR Support Unit, Canada; 3 IWK Health Centre, Canada Background: There are at least 26 systematic reviews published every day, answering the increased demand for synthesis of health research produced around the world. To produce quality research, the authors of these reviews needed to learn the appropriate methods for conducting and reporting their work. In addition to mentoring and in-person training, reviewers can learn the required skills using instructional tools and resources that are available online, including e-books, video tutorials, web-based methods guides, and webinars. Many of these resources have the advantage of being asynchronously available; they can be accessed at the point of need and can be selected for the specific experience and requirements of the learner. Objectives: To compile a database of online training resources for systematic review methods, from which we will identify common and review author-preferred elements to determine best practices in format, content, and delivery. Methods: We will conduct an environmental scan to locate online instructional supports for systematic review authors by searching the grey literature and online presences of English-language health research organizations and academic centres. We will extract and analyze data regarding the elements present in the resources to identify common practices. For example, information will be collected about the resource creator, purpose, audience, learning objectives, pedagogical approach, format, means of access, and content. These elements will be presented to a focus group of prospective and current systematic review authors for input on preferences regarding delivery methods and pedagogical style. Results: The results of the environmental scan will be presented and the key characteristics of the training resources highlighted. Conclusions: Publishers of systematic reviews, including Cochrane, have a vested interest in prospective authors acquiring the knowledge and skills to produce high-quality systematic reviews. This research will compile the existing resources available to learners and will identify elements that support the learning objectives of systematic review authors. RO 14.7 Building capacity to support evidence-informed public health: an innovative knowledge broker mentoring program Dobbins M1 , Ciliksa D2 , Yost J1 1 McMaster University, Canada; 2 The National Collaborating Centre for Methods and Tools, Canada Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Background: To achieve evidence-informed decision making (EIDM) in public health requires both organizational support and commitment alongside the development of knowledge, skill and capacity among individual practitioners. One promising strategy for EIDM is knowledge brokering. Objectives: 1. Assist health departments in identifying and developing organizational capacity for EIDM. 2. Build individual capacity for EIDM among four to six public health professionals at five health departments. Methods: The senior management team of each participating health department participated in a three-hour focus group to assess organizational capacity for EIDM. During the assessment, priorities for supporting EIDM at the organizational level were identified and strategies were developed to address these priorities. The senior management teams identified six individuals to participate in a 16-month knowledge broker mentoring program where participants receive intense training in all steps in the EIDM process. The program consists of face-to-face interactive workshops at McMaster University (10 days in total), monthly webinars, and biweekly follow-up teleconferences with an expert in EIDM. Individual knowledge and skill is being evaluated at baseline and will be evaluated again at completion of the program along with another organizational assessment. Results: Initial feedback from the first face-to-face workshop was overwhelmingly positive. All organizations are developing strategies to integrate EIDM into routine decision making. Each health department has allocated resources to cover the costs of their staff to attend the workshops, as well as the cost to deliver the workshops, and has committed two hours per week per staff member to continue developing their EIDM knowledge and skills throughout the program. Conclusions: This program will address an priority area in public health by assisting health departments to develop and support evidence-informed decision making throughout the organization. RO 14.8 A program to evaluate Cochrane training for authors of systematic reviews Cumpston M1 , Page M2 1 Cochrane Central Executive Team, Australia; 2 Cochrane Australia, United Kingdom Background: Although Cochrane has conducted training in systematic review methods since its inception, these activities have undergone only superficial evaluation based on attendance rates or participant satisfaction. The Cochrane Central Executive is initiating a new program to test our success in conveying the skills and knowledge required to complete a Cochrane review more rigorously, and to contribute to the body of research in the teaching of evidence-based medicine and research methods. Objectives: To identify a tool to assess skills and knowledge in conducting systematic reviews, and evaluate the effectiveness of Cochrane training in systematic review methods. Methods: An assessment tool to assess the skills and knowledge required to conduct systematic reviews will be identified, Cochrane Database Syst Rev Suppl 1–327 (2015) 91 and any necessary adaptation of the tool completed. As it is likely that the tool will have been developed for a different purpose (such as appraisal or reporting of systematic reviews), the tool will be validated with a diverse sample of authors to ensure construct validity and discrimination between different levels of expertise. A standardised program of evaluation will then be implemented across the Cochrane Trainers’ Network. This will enable the tool to be tested across a much larger sample, including participants in face-to-face and online training, while providing access for the first time to a data set assessing the skills and knowledge of trainee Cochrane authors. Additional components of the evaluation program will include a review of Cochrane Training materials to ensure alignment with core skills and knowledge; an assessment of first submitted draft manuscripts to assess implementation of skills and knowledge in practice; and an audit of equity of participation in training by geographically and linguistically diverse authors. Results: Preliminary results and detailed project plans for this evaluation program will be presented. Conclusions: We anticipate that progress towards a meaningful evaluation of the effectiveness of Cochrane training activities will benefit both authors and trainers, and ultimately improve the quality of Cochrane systematic reviews. Rapid oral session 15 Implementation and qualitative synthesis RO 15.1 How are realist reviews conducted? A systematic mapping review of realist reviews Berg R1 , Nanavati J2 1 Norwegian Knowledge Center for the Health Services, Norway; 2 Lakeside Upper School, USA Background: A realist review is an interpretive theory-driven narrative summary of findings from primary studies of complex social interventions. Such a review provides an explanatory analysis aimed at discerning what works for whom, in what circumstances, in what respects, and how. Objectives: To describe, with regard to realist review, how much has been done, in what areas, and what are the methodological characteristics of published realist reviews. Methods: We used a systematic mapping review approach, which captures and describes the literature in one specific field of study. We conducted systematic searches in seven international databases from January 2004 to January 2015, examined relevant websites and listserves, and contacted experts. Study selection was performed by two reviewers independently and data extraction was performed by one reviewer and checked by a second. For each included review, details of the review characteristics and methodological process were extracted (53 variables), and descriptive analyses were performed. Results: We included 54 realist Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. reviews (73 publications), conducted primarily by British researchers. Three-quarters of the titles identified the study as a realist review or realist synthesis. About half (46%) of the reviews explained the selection and appraisal of documents, and a quarter (24%) explained the data extraction process. Twenty-nine of the reviews (54%) used a data extraction form. The processes of selecting documents, appraisal of evidence, and data extraction were rarely done in duplicate (13% to 22%). About half (54%) provided a document flow diagram and 39% provided information on the characteristics of the documents included in the review. The range of documents included was six to 276 (it was unclear in five). Conclusions: In ten years, there has been an exponential growth of published realist reviews from a broad spectrum of health and related research areas. The reviews vary greatly in terms of purpose, methodological rigor, and quality of reporting. There may be a need for methodological standardization to ensure the utility and strength of its evidence. RO 15.2 Development of a checklist to assess implementation (Ch-IMP) in systematic reviews: the case of provider-based programs targeting children and youth Cargo M1 , Stankov I1 , Thomas J2 , Saini M3 , Rogers P4 , Mayo-Wilson E5 , Hannes K6 1 University of South Australia, Australia; 2 UCL Institute of Education, University College London, United Kingdom; 3 University of Toronto, Canada; 4 RMIT University (Royal Melbourne Institute of Technology), Australia; 5 Johns Hopkins Bloomberg School of Public Health, USA; 6 KU Leuven, Belgium Background: Information on implementation is required in systematic effectiveness reviews to facilitate the translation and uptake of evidence. To capture whether and how implementation is assessed in reviews a checklist for implementation (Ch-IMP) was developed and piloted in a cohort of systematic reviews on provider-based prevention and treatment interventions targeting children and youth. Objectives: This presentation reports on the inter-rater reliability and feasibility of the Ch-IMP and outlines reasons for discrepant ratings. Methods: Checklist domains were informed by a framework for program theory. Items in four domains (i.e. environment, process evaluation, action model, change model) were generated from a literature review. The checklist was pilot-tested on 27 reviews targeting children and youths. Two raters independently extracted information on 47 items, which included fidelity, dose and reach. Inter-rater reliability was evaluated using percentage agreement and kappa coefficients. Reasons for discrepant ratings were content analysed. Results: Kappa coefficients ranged from 0.37 to 1.00 and were not influenced by one-sided bias. Most kappa values were classified as excellent (n = 20) or good (n = 17) with a few items categorised as fair (n = 7) or poor (n = 1). Prevalence-adjusted kappa coefficients indicate good or excellent agreement for all but Cochrane Database Syst Rev Suppl 1–327 (2015) 92 Figure 1 one item. Four areas contributed to scoring discrepancies: (1) clarity or sufficiency of information provided in the review; (2) information missed in the review; (3) issues encountered with the tool; and (4) issues encountered at the review-level. Use of the tool demands a time investment. As such, adjustment is required to improve its feasibility for wider use. Conclusions: Results suggest that the Ch-IMP is a promising checklist for assessing whether reviews of provider-based programs targeting children and youth consider the impact of implementation variables. Used by authors and editors, the checklist may improve the quality of systematic reviews. Furthermore, the checklist shows promise as a pedagogical tool to facilitate the extraction and reporting of implementation characteristics. RO 15.3 Mixed studies reviews: types of synthesis designs used to combine qualitative and quantitative evidence Hong QN1 , Pluye P1 , Wassef M2 , Bujold M1 1 McGill University, Canada; 2 Institut national de santé publique du Québec (INSPQ), Canada Background: Reviewing qualitative and quantitative evidence has been advocated to provide in-depth answers to review questions involving complex phenomena. This type of review is emerging, and several synthesis designs have been proposed. Objective: Using a mixed methods research framework, we aim to identify and describe the different types of synthesis designs applied in systematic mixed studies reviews (SMSR). Methods: SMSR were searched for in six databases (MEDLINE, PsycINFO, Embase, CINAHL, AMED, and Web of Science) from inception through December 2014. Academic journal papers were selected for inclusion if they were systematic reviews combining qualitative, quantitative and/or mixed methods studies. Results: A total of 7006 records were found, 470 of which met our Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. eligibility criteria. The number of published systematic reviews combining qualitative and quantitative evidence has increased considerably over the past decade; passing from less than 15 in 2006 to more than 100 in 2014 (Figure 1). Three main types of synthesis designs were identified. A first type is convergent synthesis designs, in which results of included studies are transformed into either qualitative (e.g. themes) or quantitative (e.g. statistics) results. A second type is sequential synthesis designs, in which two synthesis methods are used and the results of both syntheses are compared and integrated: a quantitative synthesis informs the qualitative synthesis (sequential exploratory synthesis design) or a qualitative synthesis informs the quantitative synthesis (sequential explanatory synthesis design). A third design, named parallel, consists in two independent syntheses of qualitative and quantitative data without integration of the results of both syntheses. Conclusion: Using the mixed methods research framework as a starting point, this review describes the main types of synthesis designs usually applied in SMSR for combining qualitative and quantitative evidence. Results show that these reviews are becoming popular and warrant the need for additional methodological development. RO 15.4 Adults with kidney disease and dietary recommendations: integration of quantitative and qualitative evidence to involve patient perspectives Palmer S1 , Soroka K1 , Hanson C2 , Craig J1 , Ruospo M3 , Campbell K4 , Strippoli G1 , Tong A2 1 Cochrane Renal Group, Australia; 2 Centre for Kidney Research, Australia; 3 University of Eastern Piedmont, Italy; 4 University of Queensland, Australia Background: Patients rank understanding the effectiveness of lifestyle programs, including diet, as their top treatment uncertainty in the treatment of kidney disease. However, Cochrane Database Syst Rev Suppl 1–327 (2015) 93 existing research about the effects of specific dietary recommendations in kidney disease has not been summarized and does not account for patient experiences. Objectives: To summarize effects of dietary advice among adults with kidney disease together with patient experiences to generate recommendations for research. Methods: We conducted a systematic review of randomized controlled trials (RCTs) of any dietary intervention in chronic kidney disease and meta-analysis of treatment effects. Concurrently we did a thematic synthesis of qualitative studies of patient perceptions and experiences of dietary management. The two reviews were combined in a cross-study synthesis to identify potential dietary interventions informed by patient perspectives. Results: Seventy-eight RCTs (5945 patients) with kidney disease evaluated dietary interventions including protein intake and type, olive oil supplements, and counselling. Outcome data were sparse and effects of dietary approaches on survival and quality of life were very uncertain. From 46 qualitative studies on the experiences of dietary restrictions involving 816 patients, we identified five themes. Patients reported that dietary interventions interfered with relationships, required navigation, led to constant fighting against temptation, allowed them the opportunity to optimize their health, and increased patient empowerment in healthcare. Cross-study synthesis showed that dietary recommendations are experienced intensely by patients, but issues such as social acceptance, self-efficacy, and navigating change are not considered within existing trials. Conclusions: Our review identified a sparse evidence-base for dietary recommendations among people with kidney disease, which is grossly mismatched to the intense burden and importance of dietary modifications experienced by patients. Interventions that address a patient-centered research agenda including dietary implementation, social disruption and treatment effectiveness are needed. effectiveness are needed. RO 15.5 Applying framework synthesis to understand complexity in systematic reviews: a methodological systematic review Brunton G1 , Oliver S1 , Thomas J1 1 EPPI-Centre, UCL IOE London, United Kingdom Background: The amount of data extracted from primary studies during a complex systematic review can create challenges for synthesis, interpretation, and subsequent knowledge exchange. ‘Black box’ reviews that minimise data do not cope easily with, nor address adequately, heterogeneous interventions and populations. Systematic reviews increasingly employ framework synthesis to address these issues. However, some differences exist in how methods, outputs and underlying ways of knowing are described. It is thus timely to examine the evolving use of framework synthesis in systematic reviewing. Objectives: This systematic review aimed to address the following Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. research questions: How do methods of framework synthesis compare across reviews? Where is framework synthesis located within a range of research synthesis methods (e.g. meta-ethnography, meta-analysis)? What problems are addressed specifically by framework synthesis? Methods: Reference searches were undertaken via ASSIA, PsycInfo, PubMed, and Web of Science, key contacts and reference lists. Included reports described and/or reflected on the use of framework synthesis in systematic reviews. Data addressing the research questions were extracted. Report characteristics were compared and contrasted using thematic analysis. Results: We included 15 out of 144 citations located. Seven reports reflected on the applied use of framework synthesis; four described specific details to apply the method; and four placed it within a range of other synthesis methods. The methods of theory selection and development varied, and a deductive but iterative approach was used most often across studies. In terms of epistemological position, most reports placed the method between critical and scientific realism. Conclusions: Use of an a priori framework, iterative coding and thematic development indicate that framework synthesis is deductive; but its practical use in engaging stakeholders places the method between critical and scientific realism. Methods of selecting the a priori framework and use as a stakeholder engagement decision tool are developing; these could fill gaps not yet addressed by other methods. RO 15.6 Complex interventions and evidence translation to specific contexts: how additional analytic approaches can meet the needs of decision-makers Sutcliffe K1 , Richardson M1 , Rees R1 , Thomas J1 1 EPPI Centre, Institute of Education, University College London, United Kingdom Background: Identifying the ‘active ingredients’ of complex interventions is challenging but essential for practical implementation of review findings; some intervention approaches may be appropriate and effective in some contexts, but fail in others. For example, Weight Management Programmes (WMPs) are complex multicomponent interventions, and whilst we know that, overall, WMPs incorporating both diet and exercise components are more effective than those addressing diet or exercise alone, attempts to explain the substantial residual variance in outcomes have been unsuccessful. This equivocal picture therefore cannot help decision-makers determine the detail of which interventions to favour in which situations. Objectives: In a review for policy makers and WMP commissioners in the UK we sought to reveal sufficiently fine-grained evidence about the characteristics of successful WMPs to enable this evidence to be used in decision-making. Methods: Firstly, in addition to identifying intervention characteristics from trial reports, we conducted a synthesis of qualitative research to identify WMP characteristics based on users’ and providers’ Cochrane Database Syst Rev Suppl 1–327 (2015) 94 experiences. We then coded descriptions of evaluated WMPs according to how they matched the views of WMP users and providers. Secondly, to quantify the extent to which these WMP characteristics were associated with trial outcomes, we undertook a qualitative comparative analysis (QCA). QCA can handle situations where there are relatively small numbers of studies and multiple pathways to success better than meta-regression. Thirdly, to enhance the utility of the review, we consulted with a small number of local authorities and service commissioners on the nature and range of WMPs they currently provide and the appropriate presentation of findings. Conclusion: The translation of research knowledge into policy and practice is essential if systematic reviews are to be more than an academic exercise. New approaches for translation work may be necessary for examining complex social interventions. of the study in both sample size calculation and statistical analysis. Information on general quality criteria such as cluster consent, trial registration or publication of study protocol was also reported more rigorously in those studies. Conclusions: We found that two-thirds of c-RCTs dealing with complex interventions in a primary care setting were unable to provide evidence of improvements in patient care. This has to be kept in mind when designing future studies aimed at improving the current health care system. RO 15.8 Challenges conducting a Cochrane Review of public health ‘implementation strategies’ Williams C1 , Wolfenden L2 Hunter Medical Research Institute, The University of Newcastle, Australia; 2 The University of Newcastle, Australia 1 RO 15.7 Is effectiveness of complex interventions in primary care in cluster-randomized trials linked to methodological rigor? A systematic review Pregartner G1 , Siebenhofer A2 , Erckenbrecht S3 , Berghold A1 1 Institute for Medical Informatics, Statistics and Documentation, Medical University of Graz, Austria; 2 Institute of General Practice and Health Services Research, Medical University of Graz, Austria & Institute of General Practice, Goethe University Frankfurt, Germany; 3 AQUA-Institute for Applied Quality Improvement and Research in Health Care, Goettingen, Germany Background: In primary care, most interventions aimed at improving patient care involve those providing it and are complex in nature. Cluster-randomized controlled trials (c-RCTs) are common because they take into account the special structure of these studies. However, they also require a lot of effort to design and conduct. Objectives: The aim of this study was thus to assess how often c-RCTs dealing with complex interventions were able to show significant improvements over routine care. Furthermore, we investigated potential quality differences between studies showing an effect and those that did not. Methods: We searched MEDLINE and the Cochrane Database of Systematic Reviews for c-RCTs published in eight journals (BMJ, British Journal of General Practice, Family Practice, Preventive Medicine, Annals of Internal Medicine, Journal of General Internal Medicine, Paediatrics, Canadian Medical Journal) from 1946 to April 2014. We considered c-RCTs of complex interventions in general practices that had at least one-year follow-up and included a patient-relevant primary outcome. For each study, we assessed the effectiveness of its intervention and the reporting of 18 quality criteria. Results: We found 21 papers that fulfilled our inclusion criteria, seven of which (33%) showed an intervention effect for at least one primary outcome; four showed an effect for some but not all primary outcomes; and three for all. The latter studies all accounted for the clustered structure Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Background: Targeting health risk behaviours in children is an important aspect of chronic disease prevention, as heath behaviours established in childhood are likely to track into adulthood. Schools are an attractive setting for the implementation of child-focused chronic disease prevention interventions. However, many recommended prevention interventions are not adopted or implemented by schools, despite government mandates to do so. Studying the effectiveness of strategies designed to improve implementation, and why these strategies succeed or fail, provides critical information for future improved adoption and use of recommended practices. We are conducting two systematic reviews to assess the effectiveness of implementation strategies for school-based chronic disease prevention policies, programs or practices (interventions). Objectives: To describe and discuss conceptual and practical challenges encountered in conducting an implementation review for public health. Methods: Data concerning the conceptual and practical difficulties encountered in the conduct of the reviews has been systematically captured and will be presented as a case study. Results: The reviews are currently underway. The key challenge encountered has been distinguishing ‘intervention strategies to support implementation’ from the ‘intervention’ to be implemented. The issue is exemplified by whether health behaviour interventions, which do not primarily aim to influence implementation constitute implementation. Undoubtedly, the implementation strategy and the ‘intervention’ should be considered when interpreting our findings as they both have an effect on outcome. However, the naming function for the ’intervention’ in the Cochrane framework potentially adds further complexity to this situation. Conclusions: An extensive account of these conceptual and practical challenges will be presented at the 2015 Cochrane Colloquium. Cochrane Database Syst Rev Suppl 1–327 (2015) 95 RO 15.9 Making sense of complex interventions: evidence synthesis through harvest plots and post-review expert consultations Burns J1 , Polus S1 , Brereton L2 , Chilcott J2 , Ward S2 , Rehfuess E1 1 University of Munich, Germany; 2 ScHARR, University of Sheffield, United Kingdom Background: Systematic reviews of complex interventions frequently collect and synthesize clinically and methodologically heterogeneous evidence. In such cases, authors often deem meta-analysis inappropriate and instead perform a narrative synthesis. Other forms of synthesis in such instances are currently understudied. Objectives: To employ harvest plots to assess the effectiveness of a complex palliative care intervention, and to use gap analysis and post-review expert stakeholder consultations to make sense of the heterogeneous evidence base. Methods: We updated a review of home-based palliative care services, originally by Gomes et al 2013. Due to expected heterogeneity, we decided a priori to forego meta-analysis. We synthesized evidence using harvest plots, allowing us to visualize trends in overall effectiveness, and among subgroups. Interventions focusing on informal caregivers (which we refer to as reinforced home-based services), were of particular interest. In a subsequent gap analysis, we identified issues either not addressed through, or arising as a result of the review. With the goal of exploring and explaining the results from the effectiveness assessment further, we discussed these gaps with experts in designing, implementing and evaluating palliative care services (n = 8). Results: We included 10 new studies (total n = 30). Harvest plots showed a mix of mostly positive and neutral effects across patient and caregiver outcomes, including pain, symptom control, quality of life, psychological health, death at home, hospitalization, coping, and satisfaction with care. Reinforced home-based models, however, showed little positive effect for caregivers. Research gaps discussed with experts include potential active components, timing of palliative care, and possible reasons for the ineffectiveness of reinforced services. Conclusions: The use of harvest plots facilitated the synthesis and clear presentation of a very heterogeneous evidence base. The subsequent gap analysis and stakeholder consultations with palliative care experts allowed further exploration of the evidence, leading to more meaningful and usable conclusions for decision-makers. Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Rapid oral session 16 Rapid reviews RO 16.1 ‘Rapid review’ methodology to inform a policy decision: when is it fit for purpose? A case study Llewellyn A1 , Simmonds M1 , Brunton G2 , Thomas J2 , Sowden A1 1 Centre for Reviews and Dissemination, University of York, United Kingdom; 2 EPPI-Centre, Social Science Research Unit, UCL Institute of Education, United Kingdom Background: There appears to be an increasing use of ‘rapid reviews’ to inform policy decisions in a timely manner. However, there is relatively little consensus on rapid review methodology. Objectives: To present as a case study our experience of conducting a ‘rapid review’ to inform a policy decision, and to discuss the strengths and limitations of our approach. Methods: A systematic review of the impact of antiretroviral therapy (ART) on liver disease progression to inform policy decisions relating to patients co-infected with HIV and chronic hepatitis C was commissioned by the Department of Health in England. The short timeline (10 weeks) for completion of the review meant that a priori decisions were made to restrict: (1) the number of sources searched; (2) duplication in study selection; (3) patient and public consultation. Results: We searched MEDLINE and EMBASE and identified 1748 potentially relevant studies. Thirteen observational studies were included in the review, of which six were combined in a meta-analysis. We found that lack of randomised controlled trial evidence is more likely to have a greater impact on the conclusions drawn than weaknesses in review methods. We will discuss specific issues encountered during the review, notably involvement of different stakeholders, ‘trust’ in rapid review findings, and willingness of decision makers to potentially compromise methodological rigour. Conclusions: Our review provided a concise and transparent picture of the evidence available to inform future policy decisions related to patients co-infected with HIV and chronic hepatitis C. It was produced within a short time frame and omitted elements of the systematic review process such as comprehensive searching and double screening, and was therefore potentially at higher risk of bias and error than reviews adopting ‘gold standard’ methods. Implications of the review methodology used, the nature of the evidence synthesised and the policy context will be discussed. Cochrane Database Syst Rev Suppl 1–327 (2015) 96 RO 16.2 Streamlining systematic reviews to provide healthcare decision-makers with more timely evidence: the DERP experience Peterson K1 , Selph S1 , Holmes R1 , Feltner C2 , Jonas D2 , Holzhammer B1 , McDonagh M1 1 Pacific Northwest Evidence-Based Practice Center, USA; 2 Department of Medicine, University of North Carolina-Chapel Hill, USA Background: High-quality systematic reviews are the gold standard of evidence synthesis, but can be time-prohibitive for healthcare decision-makers faced with shorter-term information needs. In an effort to better provide Drug Effectiveness Review Project (DERP) decision-makers with the right information at the right time, in 2012 the Pacific Northwest Evidence-based Practice Center (PNW EPC) started streamlining their standard systematic review processes, reducing timelines for new reviews by four months and by three months for updates. The PNW EPC also created two new abbreviated products, single drug reviews and reviews of reviews. Objective: Describe the DERP streamlining process and its impact on quality of reports produced between July 2012 and October 2015. Methods: For each streamlined DERP report, two authors completed a questionnaire that assessed: (1) types of scope and methodological modifications; (2) impact on report quality indicators and conclusions; (3) workload manageability; and (4) suggestions for improvement. We used a mixture of question formats, including yes/no, multiple choice, Likert scale ratings and open-ended responses. We will use descriptive statistics and narrative methods to summarize author responses. Results: We assessed six new and four updated DERP reports. Streamlining included scope restrictions, abbreviated data abstraction, less dual review of evidence strength ratings, and less meta-analysis. For report quality, authors indicated no change to key finding clarity, organization, applicability, or risk of bias, but reported slightly worse limitations and workload manageability. Due to removal of indirect and noncomparative data, streamlined updates resulted in more evidence gaps. Our findings are potentially limited by recall bias due to retrospective data collection, and by being specific to drug-drug comparisons. Conclusions: In order to provide health care decision-makers with more timely evidence, it is possible to streamline the standard systematic review process with limited impact on author-assessed report quality. Further assessment of report quality is needed from external reviewers and users. RO 16.3 Rapid reviews to inform state health policy decisions King V1 , Gerrity M1 1 OHSU, USA Background: Medicaid, a US federal-state partnership, provides care to poor and disabled children and adults. Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Seventeen state Medicaid programs collaborate, as the Medicaid Evidence-based Decisions (MED) project, with the Center for Evidence-based Policy (CEBP) to produce rapid reviews (RR) to inform health policy decisions. Objectives: To describe the methods used to produce RRs for US state policymakers. Methods: Review of CEBP RR methods, their evolution, and current outputs. Results: Between 2006 and 2015 the CEBP produced over 300 RRs. Initial MED evidence reports (2006-2008) were brief evidence summaries of SRs from trusted core sources such as the Cochrane Library. Reports sometimes included searches to update SRs, but critical appraisal (CA) of studies was not routine. In 2008, participating policymakers requested more complete evidence reviews, and wanted the option to add clinical practice guidelines, with CA of the guidelines, and payer coverage policies. Since 2009, CEBP has routinely incorporated the following methodology elements into RR production: refinement of topic scope and protocol with nominating state; multiple database search for SRs and update searches using PubMed; single reviewer study selection/abstraction/CA, with second reviewer validation; narrative synthesis; internal review; and proprietary publication. Optional elements, incorporated based on the topic and needs of the nominator, include: peer-reviewed search strategy; search of additional databases; industry dossiers; guidelines and payer policies; dual review of selection/abstraction/CA; external review and public dissemination. Production timeline is generally three to four months from protocol agreement. Conclusions: Medicaid programs have successfully collaborated with the CEBP over the last decade to develop RRs to inform policy decisions. Evidence reviews alone seldom provide sufficient information for good policy formation. However, critical analyses of evidence and appraisal of the quality and concordance of guidelines and policies with that evidence can assist policymakers in making evidence-informed decisions within relevant timeframes and available resources. RO 16.4 If a rapid review is the solution, what is the problem? Mann M1 , Naughton A2 , Maguire S1 , Kemp A1 , Cowley L1 1 Cardiff University, United Kingdom; 2 Public Health Wales, United Kingdom Background: Rapid reviews have emerged as an efficient approach to synthesizing evidence. There is a growing number of rapid review ’methods’ nevertheless; there is not an accepted standardised methodology. Consequently, we used systematic review methodology to conduct a rapid review to identify evidence behind the self-reported features in adolescents who are experiencing neglect and/or emotional maltreatment. Objectives: To identify the methodological challenges related to the conducting a rapid review using streamlined systematic review methods. Methods: We used a focused search for all types of studies published in the UK Cochrane Database Syst Rev Suppl 1–327 (2015) 97 from 1990 to 2014. Due to the topic area, we searched a wide range of databases relating to health, medicine, education and psychology. Two reviewers independently screened the papers to determine relevance according to eligibility criteria, followed by full data extraction and critical appraisal by one reviewer, with independent verification by a second reviewer. Results: The search retrieved 4388 studies and 2568 records were screened after manual de-duplication and removing clearly irrelevant studies. Two reviewers independently evaluated 279 studies for quality and eligibility; subsequently 19 papers were included in the review. Conclusions: Drawing upon lessons learned from our review, we identify challenges encountered at each stage of the review process and discuss the solutions identified. RO 16.5 Rapid review on the effectiveness of personal protective equipment for healthcare workers caring for patients with filovirus disease Hersi M1 , Stevens A2 , Quach P1 , Hamel C1 , Thavorn K3 , Garritty C2 , Skidmore B4 , Vallenas C5 , Norris SL5 , Egger M6 , Eremin S5 , Ferri M7 , Shindo N5 , Moher D3 1 Ottawa Hospital Research Institute, Canada; 2 Ottawa Hospital Research Institute, Canada; 3 Ottawa Hospital Research Institute, University of Ottawa, Canada; 4 Independent Consultant, Canada; 5 World Health Organization, Switzerland; 6 University of Bern, University of Cape Town, Switzerland/South Africa; 7 University of Calgary, Canada Background: The West African outbreak of the Ebola virus disease in 2014 evolved rapidly, yielding the highest number of cases and deaths of outbreaks to date. We performed a rapid review of the evidence to inform recommendations issued by the World Health Organization on the use of personal protective equipment (PPE) by healthcare workers managing patients with known or suspected filovirus (Ebola and Marburg) disease. Objectives: To determine the comparative benefits and harms of various PPE (e.g. double gloves, full face protection) components. Methods: A rapid review (accelerated and/or modified systematic review methods) guided by a protocol was conducted over seven weeks. Bibliographic databases, grey literature sources, and supplemental sources were searched. Eligibility criteria initially included only comparative studies on Ebola and Marburg diseases reported in English or French, but criteria were expanded to studies on other, viral hemorrhagic fevers and to non-comparative designs because comparative studies were lacking. Titles and abstracts were reviewed by one person, and a second person verified potentially excluded records. Full-text articles were reviewed by two independent people. Meta-analysis of data was not done, but plots summarizing data were produced where appropriate. The domains of the Grading of Recommendations Assessment Development and Evaluation framework (GRADE) were used to inform judgments on the quality of the evidence. Results: No comparative studies were located. Thirty non-comparative (eight related to Ebola) studies were included; 27 studies provided data on viral transmission, while nine studies Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. reported on other adverse events, such as needle-stick injuries. The quality of the body of evidence for all outcomes was low. In general, studies reported information on PPE components and infection prevention and control protocols poorly. Conclusions: Insufficient evidence exists to draw conclusions on the comparative effectiveness of various types of PPE. Additional research is urgently needed, and considerations exist for future research. RO 16.6 Rapid reviews to inform Ebola preparedness efforts at hospitals Mitchell M1 , Price D1 , Umscheid C1 , Mull N1 1 University of Pennsylvania Health System, USA Background: US hospitals are preparing to evaluate and treat returning travelers at risk for Ebola Virus Disease (EVD). Our hospital is one of 35 designated by the Centers for Disease Control and Prevention (CDC) as an Ebola treatment center. Objectives: Acquire, appraise, and apply best available evidence to develop protocols rapidly for the treatment of patients with possible EVD. Methods: The focus and scope of our reports were determined in an iterative process between the research analyst, the clinicians, and the center director. Rapid review products from the center include evidence inventory reports that describe the type and quantity of evidence available in a specified area, evidence advisories based on secondary sources, and evidence reviews analyzing primary studies. Results: The initial clinical query sought evidence on the effectiveness of rehydration, anti-emetic, and anti-diarrheal treatments in patients with EVD. An advisory was completed in 14 days, finding one guideline for caregivers in rural Gabon, but no studies in EVD patients. Given the limited direct evidence, the clinicians then requested a rapid review of oral rehydration for the broader population of patients with any acute infectious gastrointestinal diseases. That review identified a strong consensus of guidelines favoring reduced-osmolarity oral rehydration solutions. Based on those guidelines, the scope of review of primary literature was narrowed to studies of adults, which helped us complete the evidence tables and provide preliminary findings 14 days after the report request, and finalize the review and incorporate into a clinical protocol six business days later. The quality of the evidence base was low. Conclusions: Close consultation between local evidence-based practice centers and clinical decision-makers can expedite evidence reviews to respond to urgent needs. RO 16.7 Possibilities of standardisation for rapid reviews –proposal for the German Federal Ministry for Education and Research Buchberger B1 , Kossmann B1 , Krabbe L1 , Mattivi JT1 1 University of Duisburg-Essen, Institute for Health Care Management and Research, Germany Cochrane Database Syst Rev Suppl 1–327 (2015) 98 Background: Through comparison of systematic reviews and rapid reviews for surgical interventions using AMSTAR, an instrument for quality assessment, we identified ambiguity concerning the term ‘rapid’ as well as a need for standardisation of methods for rapid reviews. Due to the urgency of some questions in health care, an acceleration of the review process is a high priority, and therefore, standardisation will be very helpful. There are many possibilities for methodological restrictions, but weak empirical evidence regarding the consequences such as increased potential of bias. Objectives: Our aims are to develop and discuss a standardised rapid review methodology, to conduct a rapid review based on that methodology and a systematic review at the same time and concerning an identical research question, and to compare and discuss the results further. Methods: The research project will be divided into three parts: (A) Comprehensive systematic literature searches and handsearches for existing standards for rapid reviews will be conducted on websites of Health Technology Assessment agencies, and governmental and non-governmental organisations. Based on a synopsis of the results, we will interview methodology experts and organise focus groups for discussion. The results will lead to the development of a standardised format for rapid reviews that will be revised by experts. (B) The developed format will be tested by conducting a rapid review and a systematic review at the same time on an identical question. (C) The results, assessments of included evidence, conclusions, and recommendations of both reviews will be compared and discussed in focus groups consisting of decision makers in health care. Conclusions: Comparing the results of a rapid review conducted according to a newly developed methodology with a systematic review on the same topic and in parallel design will lead to conclusions about possible methodological restrictions, their potential of bias and the time, as well as resources, saved by accelerating the review process. A standardised methodology for rapid reviews will help policy-makers with time pressures to decide swiftly with greater certainty. Rapid oral session 17 Network meta-analysis RO 17.1 Characteristics of network meta-analysis reviews from Cochrane Porfirio G1 , Nogueira G1 , Cabrera A1 , Parra M1 , Freitas C1 , Torloni MR1 , Riera R1 , Atallah A1 1 Brazilian Cochrane Centre, Brazil Background: Network meta-analysis is a study design that allows indirect comparisons between interventions when there are no studies with direct comparisons. Although there is no standard method, crucial points, as the level of similarity within studies (called transitivity or similarity) and, the analysis and results (called consistency or coherence) Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. became well known. Considering the recent growth and the lack of well-established standards for network meta-analysis, a question arises: What are the characteristics of network meta-analysis reviews from Cochrane? Objectives: To describe the characteristics of network meta-analysis reviews published by Cochrane. Methods: We performed a descriptive study at the Brazilian Cochrane Centre (BCC). We included Cochrane Systematic Reviews using the expression ’network meta-analysis’ in any part of the text. We selected the studies after full readings. The extracted outcomes were: review characteristics, related review group, year and stage of publication. Results: The search was conducted in March 2015. We identified 104 studies and excluded 71 studies after full text reading. From the 33 remaining studies, 16 were reviews, 12 protocols and 5 overviews distributed in 21 review groups; 48% of studies were published in 2014 (16/23); 36% mentioned ’network meta-analysis’ in the title. Considering only finished studies (reviews and overviews), 57% (12/21) mentioned having performed transitivity assessment and 57% mentioned that that assessed inconsistency, although they were not exclusively the same studies. Graphs and figures representing the network meta-analysis were found in only three reviews. Eight reviews mentioned the use of STATA and seven mentioned WinBUGS as the software for analysis. Conclusions: We believe that there is a need to define description standards for networks meta-analysis within Cochrane, with emphasis on transitivity and inconsistency. Standardizing graphs and figures would also be beneficial. Those were present in a minority of studies. RO 17.2 Comparing various approaches for network meta-analysis regarding effect estimation and evaluation of inconsistency Bender R1 , Sturtz S1 , Sieben W1 , Kiefer C1 1 IQWiG, Germany Background: Network meta-analysis is an important extension of pairwise meta-analysis to compare more than two interventions and to analyze indirect and direct evidence simultaneously. It is becoming more and more popular in systematic reviews and health technology assessment. Although for pairwise meta-analysis the properties of the different approaches are well examined, little is known for the approaches in network meta-analysis. Objectives: To compare methods that are currently available for estimating and assessing inconsistency in network meta-analysis. Methods: A simulation study was conducted to evaluate complex networks with up to five interventions and different patterns. We analyzed the impact of different network-sizes, different amounts of inconsistency and heterogeneity on MSE and coverage of established and new approaches. This exceeds previous simulations studies, which have been conducted by others. Methods: We found that, with a high degree of inconsistency in the network, none of the evaluated effect estimators produced reliable results. For a network with no or just moderate inconsistency the Bayesian and the frequentist estimator showed acceptable properties, whereas Cochrane Database Syst Rev Suppl 1–327 (2015) 99 the latter one showed slightly better results. We also found a dependency on the amount of heterogeneity in the network. Conclusions: Our results underline the need to assess inconsistency in network meta-analyses reliably as available measures for inconsistency may be misleading in many situations. We therefore conclude that it is also important to assess similarity (regarding population, intervention, etc.) and heterogeneity to reduce inconsistency in the network in advance. Nevertheless, effect estimators should be used which are suitable in the case of moderate inconsistency. RO 17.3 Statistical models for network meta-analysis: an empirical comparison Debray T1 , Schuit E2 , Efthimiou O3 , Jansen J4 , Ioannidis J5 , Moons K1 1 Julius Center for Health Sciences and Primary Care, The Netherlands; 2 Stanfor University, USA; 3 University of Ioannina, Greece; 4 Redwood Outcomes, USA; 5 Stanford University, USA Background: Network meta-analysis (NMA) can be conducted to compare the relative efficacy of multiple treatments. It is currently unclear whether access to individual participant data (IPD) is beneficial in NMA. Methods: We discuss four key issues in a NMA and present several statistical methods to address these issues. In particular, we describe methods to overcome baseline imbalance (e.g. due to unlucky randomization), to investigate between-study heterogeneity in treatment effect, to investigate within-study heterogeneity in treatment effect and to minimize network inconsistency. For all methods, we highlight the role of IPD, and illustate their implementation in an empirical example dataset. Results: Results indicate that when NMA do not incorporate patient- or study-level covariates, one-stage and two-stage models lead to similar estimates of relative treatment effect. However, two-stage models yielded slightly larger standard errors and substantially larger estimates of between-study heterogeneity. For all models, estimates of treatment effect gained precision and network consistency improved when adjusting for covariates. Most importantly, NMA that were based on IPD showed greater power in investigating the presence of effect modification, and helped to improve overall network consistency. Conclusions: When investigating the relative efficacy of multiple treatments, NMA that are (partially) based on IPD may help to overcome several key issues hampering the validity and interpretability of traditional NMA analyses that are solely based on published aggregate data. RO 17.4 Exploring and accounting for the impact of interventions with scarce evidence in network meta-analysis Chaimani A1 , Salanti G1 of Hygiene & Epidemiology, School of Medicine, University of Ioannina, Greece 1 Department Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Background: Previous empirical evidence has revealed that interventions evaluated in very few trials are often placed among the most effective interventions in a network meta-analysis (NMA). Although this phenomenon can be expected for recently marketed treatments, older treatments with scarce evidence could show exaggerated results due to publication bias or because they have been compared only to suboptimal alternatives. If the network is poorly connected, such treatments are informed mainly by their potentially biased direct comparisons and may rank spuriously high in the overall ranking. Objectives: To explore the impact of interventions with scarce evidence in a NMA and to estimate the relative ranking of treatments as a function of the amount of available information in the network. Methods: We developed a network meta-regression model using the contribution of each direct comparison to the network as a predictor of the relative effects. We assumed that interventions poorly connected to the rest of the network might be favored in a direct comparison. We applied the model in a network that compares eight different stents for myocardial infraction, in which results are highly affected by the direct comparison between bare-mental stents (BMS) and control. Results: Based on the cumulative ranking probabilities, the NMA model indicated BMS as the best intervention, although newer treatments are well-known to be safer. After applying our meta-regression model BMS was placed at a lower rank and results were closer to those expected from clinical practice. Conclusions: In poorly connected networks, the contribution of indirect information might be low and some interventions may appear high in ranking because they are informed primarily by their direct comparisons. Modeling the amount of available evidence for each intervention might be a useful tool to evaluate the robustness of results from NMA. How results change if suspicious comparisons are given a lower weight might also be explored via a sensitivity analysis. RO 17.5 The impact of publication bias and small study effects on the efficacy and ranking of antipsychotics Mavridis D1 , Efthimou O2 , Mavridis S3 , Salanti G2 1 Department of Primary Education, University of Ioannina, Greece; 2 Department of Hygiene and Epidemiology, University of Ioannina, Greece; 3 Department of Psychiatry and Psychotherapy, Technische Universität München, Germany Background: Evidence from placebo-control trials comparing the least effective antipsychotics is distorted by small study effects and publication bias (PB). This raises concerns about the validity of the results from network meta-analysis (NMA). Biased evidence in placebo-control comparisons may affect the treatment effects and the ranking of the antispychotics as placebo-controlled trials account for one-third of the information in the entire network. Objectives: To assess the potential impact of small study effects (SSE) and PB on Cochrane Database Syst Rev Suppl 1–327 (2015) 100 the recently estimated relative effectiveness and ranking of pharmacological treatments for schizophrenia. To suggest statistical methods for adjusting results in the presence of SSE or PB. Methods: We used a recently published network of 167 studies involving 36,871 patients and comparing the effectiveness of 15 antipsychotics and placebo. We used novel visual and statistical methods to explore whether smaller studies are associated with larger treatment effects and if the probability of publication is associated with the magnitude of effect. We conducted a NMA of the published evidence as our primary analysis and employed a sensitivity analysis considering low, moderate and severe selection bias (that correspond to the number of unpublished trials) with an aim to evaluate robustness of point estimates and ranking. We explored whether placebo-controlled and head-to-head trials are associated with different levels of PB. Results: We found that small placebo-controlled trials exaggerated slightly the efficacy of antipsychotics and PB was not unlikely in the evidence based on placebo-controlled trials; however, ranking of antipsychotics remained robust. Conclusions: The total evidence comprises many head-to-head trials that do not appear to be prone to SSE or PB and indirect evidence ‘washes-out’ some of the bias in the placebo-controlled trials. RO 17.6 Incorporating external information on between-study heterogeneity in network meta-analysis Turner R1 , Rhodes K1 , Jackson D1 , White I1 1 Medical Research Council Biostatistics Unit, United Kingdom Background: In a network meta-analysis comparing multiple treatments, between-study heterogeneity variances are often very imprecisely estimated because data are sparse, and so standard errors can be highly unstable. External evidence obtained from modelling data from the Cochrane Database of Systematic Reviews can provide informative prior distributions for heterogeneity, tailored to particular settings. Objectives: To explore and compare approaches for specifying informative priors for multiple heterogeneity variances in a network meta-analysis. Methods: If heterogeneity variances can be assumed to be equal across all pairwise comparisons of treatments, it is straightforward to construct an informative prior for the common between-study variance. Models allowing heterogeneity variances to be unequal are more realistic, however, care must be taken to ensure that the implied variance-covariance matrices remain valid. We consider two strategies for specifying informative priors for multiple heterogeneity variances: proportional relationships among the variances; or unequal heterogeneity variances with a common informative prior. Results: Appropriate prior distributions are obtained through modelling empirical data from the Cochrane Database of Systematic Reviews. The models are applied to a network meta-analysis comparing four treatments for smoking cessation. Incorporating external information on Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. heterogeneity in the equal and unequal variance models leads to smaller heterogeneity estimates, with narrower intervals. This causes changes to the odds ratios, and their 95% intervals narrow substantially. For example, the odds ratio comparing group counselling against standard care is estimated as 3.39 (95% CI 0.98 to 16.8) when using vague priors, in the unequal variances model. This changes to 3.10 (95% CI 1.60 to 6.17) when using informative priors for heterogeneity. Conclusions: Relevant prior information on heterogeneity can be incorporated into network meta-analyses, without making unrealistic assumptions. This may improve precision for estimating treatment differences. RO 17.7 Prospective use of network meta-analysis using formal statistical monitoring Nikolakopoulou A1 , Mavridis D2 , Salanti G1 1 Department of Hygiene and Epidemiology, University of Ioannina, School of Medicine, Ioannina, Greece; 2 Department of Primary Education, University of Ioannina, Ioannina, Greece Background: Pairwise and network meta-analyses (NMA) are traditionally used retrospectively to assess existing evidence. However, previous knowledge of trial results can introduce bias due to potential selective inclusion of the key components of the review question (PICO criteria). Prospective meta-analysis overcomes this limitation by requiring the identification of eligible trials before the disclosure of their results. It has been suggested that NMA can also be undertaken prospectively; this practice though has not been adopted yet. Objectives: Designing a NMA prospectively creates a multiple testing scenario. Our objective is to develop a method for the prospective design of NMA by using formal monitoring. Methods: We extend ideas of sequential monitoring of trials to random-effects NMA. We construct efficacy and futility stopping boundaries for NMA estimates and we present the sequential NMA procedure using repeated confidence intervals (RCI). When a RCI excludes the point of no effect the NMA effect is considered conclusive. Our method leads to recommendations of whether further research is required to inform NMA and for which comparisons. We illustrate the method using a network that evaluates the effectiveness of bare mental stent (BMS), coronary artery bypass (CAB) and drug eluting stent (DES) for all-cause mortality in diabetic patients. Studies are added in a chronological order to the NMA model. Results: After the inclusion of eight studies none of the NMA summary effects was conclusive. The incorporation of the 9th and the 10th trial lead to a conclusive NMA effect estimate for the comparisons ‘BMS versus DES’ and ‘CAB vs BMS’ respectively showing the relative inferiority of BMS after accounting for multiple testing. If NMA had been planned prospectively, it would have stopped with the inclusion of the 13th study when NMA effects are conclusive for all comparisons. Conclusions: Use of sequential methods in NMA can be adopted so that trials do not address already answered questions. Such a procedure would help to save resources and prevent the allocation of participants of RCTs Cochrane Database Syst Rev Suppl 1–327 (2015) 101 to treatments proved to be inferior. RO 17.8 Hierarchical network meta-analysis models accounting for variability in nodes by treatment, dosage-category and single dosage Veroniki AA1 , Tricco A1 , Del Giovane C2 , Blondal E1 , Thavorn K3 , Hutton B3 , Straus S1 1 Li Ka Shing Knowledge Institute, St. Michael’s Hospital, Toronto, Canada; 2 Department of Diagnostic, Clinical and Public Health Medicine, University of Modena and Reggio Emilia, Modena, Italy; 3 Department of Epidemiology and Community Medicine Faculty of Medicine, University of Ottawa, Canada Background: Healthcare providers, policy-makers and patients are often interested in the effect of specific treatment dosages or categories of dosage (e.g. low, moderate, high), and the treatment-effects overall. Decisions regarding network geometry (e.g. lump related dosages together into one treatment node or maintain their separation) may impact the network meta-analysis (NMA) results, and influence decision-making. Objectives: To present hierarchical NMA models accounting for effects of treatments, dosage-categories, and single dosages, providing additional insight on different heterogeneity levels. Methods: We developed three approaches accounting for the relationship between treatment and dosage, and including up to four heterogeneity levels. The first approach assumes the dosage-effects equal to their parent treatment-effects, and involves within-study and between-study heterogeneity levels within dosage. The second approach assumes exchangeable dosage-effects from the same distribution with a common mean, and incorporates an additional heterogeneity level, the between-dosage within-treatment. The third approach assumes exchangeable dosage-effects within a specific dosage-category with either fixed or random mean dosage-category-effects across treatments. The third model still considers the first two heterogeneity levels, yet the between-dosage is evaluated within dosage-category, and an extra heterogeneity level is added, the between-dosage-category within treatment. Consistency constraints on dosage-effects can also be placed in each model. Results: Results from the application of the different approaches to empirical examples, will be ready by September 2015, and will be presented at the Colloquium along with strengths and limitations of the models. Conclusions: Different approaches regarding the classification of treatments in a network may result in important variations in interpretations drawn from NMA. We suggest that researchers account for different treatment dosages in NMA, providing additional insight on heterogeneity, inconsistency, intervention ranking, and hence decision-making. Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Rapid oral session 18 Outcomes RO 18.1 Planning and reporting of quality of life outcomes in cancer trials: a retrospective cohort study Schandelmaier S1 , Amstutz A1 , Von Elm E2 , Conen K1 , Hemkens LG1 , Bucher HC1 , Guyatt GH3 , Briel M1 , Kasenda B4 1 University Hospital, Basel, Switzerland; 2 University Hospital, Lausanne, Switzerland; 3 McMaster University, Hamilton, Canada; 4 Royal Marsden Hospital London, United Kingdom Background: Information about the impact of cancer treatments on patients’ quality of life (QoL) is of paramount importance to patients, treating physicians, and policy makers. Although cancer specific and validated QoL instruments exist, reports of randomized clinical trials (RCTs) involving cancer patients often do not include QoL outcomes; instead, they typically focus on survival or tumor size as their primary outcome. Possible reasons for lack of QoL outcomes include failure to collect QoL outcomes and failure to publish collected QoL outcome data. Objectives: To investigate whether QoL outcomes were specified in protocols of industry- or investigator-initiated RCTs enrolling cancer patients, and whether they were subsequently reported. Methods: We conducted a retrospective cohort study of RCT protocols approved by six research ethics committees in Switzerland, Germany, and Canada between 2000 and 2003 and compared 173 protocols of cancer trials to 108 corresponding publications that we identified through literature searches and investigator surveys. Results: Of the 173 RCTs, 90 (52%) specified QoL outcomes in their protocol, two (1%) as primary and 88 (51%) as secondary outcomes; 35 (20%) reported QoL outcomes in a corresponding publication (four modified from the protocol), 37 (21%) remained unpublished, and 18 (10%) were published but failed to report QoL outcomes. Failure to report QoL outcomes was not significantly associated with sponsorship, sample size, single center status, or trial discontinuation. None of the 83 (48%) RCTs that did not specify a QoL outcome in their protocol subsequently reported a QoL outcome. Conclusions: About half of cancer RCTs specified QoL outcomes in their protocols. However, only 20% reported any QoL data in associated publications. Highly relevant information for decision-making is not available to patients, physicians, systematic reviewers, and policy makers. If published, QoL outcomes in cancer trials are prone to selective reporting. Cochrane Database Syst Rev Suppl 1–327 (2015) 102 RO 18.2 Missing the wood for the trees: clinically relevant outcome measures in traumatic brain injury rehabilitation, a review of reviews Viswanathan A1 , Samuelkamaleshkumar S2 , Senthilvelkumar T2 1 South Asian Cochrane Network and Centre, India; 2 Christian Medical College, India Background: The Cochrane Handbook suggests that ‘‘indirect or surrogate outcome measures are potentially misleading and should be avoided or interpreted with caution because . . . many interventions reduce the risk for a surrogate outcome but have no effect or have harmful effects on clinically relevant outcomes, and some interventions have no effect on surrogate measures but improve clinical outcomes’’. This is very relevant to the field of neurorehabilitation, where tangible clinical end points that matter to the patients, such as functional independence, return to work and community integration should be preferred over surrogate measurements such as cognition, spasticity, and muscle power. Objectives: 1. To analyze proportions of clinically relevant outcome measures versus surrogate outcomes reported in Cochrane Systematic Reviews (CSRs) on rehabilitation interventions following traumatic brain injury (TBI). 2. To analyze the proportion of individual trials in these CSRs that report clinical outcomes. Methods: The Cochrane Library was searched for CSRs on rehabilitation interventions following traumatic brain injury. Exclusion criteria were non-traumatic etiology, and pharmacological interventions. Three authors analyzed the search results. Outcomes were extracted and categorized as clinically relevant or surrogate. Results: Thirty-seven CSRs related to TBI were screened, five met the inclusion criteria: 18 out of 49 outcomes (37%) enlisted in five CSRs were clinically relevant outcomes; the rest were surrogate outcomes. A total of 17 trials were included in these reviews. Eight trials from two CSRs (47%) had reported at least one clinically relevant outcome. In the remaining three CSRs, no study reported any clinically relevant outcome. Meta-analysis was not done in any of the five CSRs. Conclusions: Less than half of the outcomes reported in CSRs on TBI rehabilitation were clinically relevant. More than half of the trials included in these reviews did not report any clinically relevant outcome measure. RO 18.3 Outcomes in Child Health: comparing outcomes evaluated in Cochrane Reviews and patient-important outcomes identified by parents using social media Hamm M1 , Fernandes R2 , Scott SD1 , Hartling L1 1 University of Alberta, Canada; 2 Cochrane Child Health Field; Cochrane Portugal; Department of Pediatrics and Clinical Pharmacology Unit, Lisbon Academic Medical Centre, Portugal Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Background: Selection of appropriate outcomes for clinical research is increasingly focused on patient-centeredness, but outcomes reported in the literature do not necessarily align with outcomes important to patients. Objectives: To compare outcomes reported in Cochrane Reviews on pediatric acute respiratory infections (ARI) with those identified by parents as important to them and their children. Methods: Parents of children who had an ARI were recruited via social media to provide input on the outcomes that are most important to them when their child is sick. We used the registry of child-relevant Cochrane Reviews maintained by the Cochrane Child Health Field to identify reviews applicable to ARI and extracted data on their reported outcomes and the frequency with which they appeared across reviews. This set of outcomes was compiled into an online survey in which parents were asked to rate each on a scale of 1–100 based on how important they considered it to be. We also asked parents to identify other important outcomes that were not included in the list. We analyzed data descriptively using a predefined outcome domain framework and compared the prominence of outcomes evaluated in Cochrane Reviews with those identified as being important by parents. Results: We identified 35 reviews relevant to pediatric ARI. In the reviews and the survey, safety was commonly endorsed, with adverse events reported 26 times across reviews, and parents ranking major side effects (mean score 86.7/100) among their most important concerns. Parents identified severe complications as their most important concern (94.5/100). Resource utilization was more prominent in reviews, with admission rates and length of stay/time to discharge ranking in the highest quartile of outcomes measured, while parents scored similar items in the 75th percentile. Social and family outcomes (e.g. quality of life) were not often reported in reviews, but made up the most frequently mentioned outcomes when parents were asked about other concerns that were not listed. Conclusions: Safety is a primary concern to parents. Integrating patient priorities will ensure the relevance of research results. RO 18.4 A map of outcomes in trials of communication interventions for childhood vaccination: measuring too few concepts in too many ways Kaufman J1 , Ryan R1 , Bosch-Capblanch X2 , Cartier Y3 , Cliff J4 , Glenton C5 , Lewin S6 , Rada G7 , Ames H5 , Muloliwa AM8 , Oku A9 , Oyo-Ita A9 , Hill S1 1 Centre for Health Communication and Participation, La Trobe University, Australia; 2 Swiss Tropical and Public Health Institute, University of Basel, Switzerland; 3 International Union for Health Promotion and Education, France; 4 Universidade Eduardo Mondlane, Mozambique; 5 Global Health Unit, Norwegian Knowledge Centre for the Health Services, Norway; 6 Global Health Unit, Norwegian Knowledge Centre for the Health Services, Norway/Health Systems Research Unit, South African Medical Research Council, South Africa; 7 Catholic University of Chile, Chile; 8 Provincial Directorate of Health, Mozambique; 9 University of Calabar, Nigeria Cochrane Database Syst Rev Suppl 1–327 (2015) 103 Background: Core outcome sets (COSs) bring consistency to outcome measurement and strengthen systematic reviews. Most COSs focus on clinical interventions for health conditions; few address preventive health interventions.Communication for childhood vaccination is key to improving uptake and can take many forms. Measurement requires appropriate outcomes, but it is not clear what these outcomes should be or how COS methods apply to this context. Here we describe novel methods for developing a Trial Outcomes Map (TOM), a helpful tool for the early stages of COS development. Objectives: To catalogue which outcomes have been measured in childhood vaccination communication trials and describe (in)consistencies and patterns in these with a TOM. Methods: We searched the Cochrane Central Register of Controlled Trials (CENTRAL) for relevant trials and extracted outcome details. Outcomes were inductively coded into thematic groups, forming the map. Consistencies, gaps and trends were analysed. Results: We extracted data on 209 outcomes from 112 trials. We found variation and gaps in reported outcomes.Thematic grouping identified three overarching categories: vaccination-, consumer- and health system-related outcomes. Vaccination-related outcomes were reported most frequently: 80% of trials measured at least one such outcome, although these outcomes were defined and measured in a wide range of ways. Consumer-related outcomes, tied mainly to knowledge, attitudes or beliefs, were measured by 28% of trials. Health system-related outcomes were measured in 20% of trials. Trials published after 2000 appeared to include more consumer-related outcomes, possibly reflecting increased interest in shared decision making. Conclusions: This map highlights variation in the measurement of key vaccination outcomes, leading to difficulties in interpretation and synthesis. A trend towards more consumer-related outcomes emerges, but intermediate consumer- and health system-related outcomes that clarify vaccination communication causal pathways are not measured sufficiently frequently in trials. This map is a tool for trialists and reviewers and will inform the next stages of COS development. RO 18.5 Development of a core outcome set for infant colic Steutel NF1 , Korterink JJ2 , Benninga MA2 , Langendam MW3 , Tabbers MM2 1 Dept Clinical Epidemiology, Biostatistics and Bioinformatics and Dept Pediatric Gastroenterology and Nutrition, Academic Medical Centre, The Netherlands; 2 Dept Pediatric Gastroenterology and Nutrition, AcademicMedicalCentre, TheNetherlands; 3 DeptClinical Epidemiology, Biostatistics and Bioinformatics, Academic Medical Centre, The Netherlands Background and problem with a This self-limiting consequences. objectives: Infant colic (IC) is a common worldwide prevalence of 5% to 25%. disorder can have negative long-term So far, its etiology remains unknown, Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. resulting in a wide variety in interventions. A previous study showed heterogeneous outcome measures are used across therapeutic trials of IC. To facilitate and improve evidence synthesis, development of a core outcome set (COS) is necessary. Methods: In 2014, 133 health care professionals (HCPs) participated in a survey that investigated which outcome measures they used to guide their decision making when treating IC. Every outcome that was mentioned by 10% or more of the respondents was carried forward to Phase 2. Of the respondents to Phase 1, 68 (51%) agreed to complete a second questionnaire. HCPs were then asked to rate the outcomes and to select the five most relevant outcomes, for both an inpatient and outpatient setting. Results (preliminary): To date 36 (53%) HCPs have completed the second questionnaire. Duration of crying, reduced family stress, sleeping time of the infant and discomfort of the infant were considered to be most important in both outpatient and inpatient setting. In the outpatient setting quality of life was rated as important, and in the inpatient setting reduced hospital admission (and/or duration). Conclusions: These preliminary results show that there is not much difference in outcome measures that guide treatment decisions in an outpatient and inpatient setting for IC. Furthermore, the results appear to be quite homogeneous across respondents. In a later phase of the project parental views will be taken into account to complete this COS. This COS may enable researchers to standardize the outcomes they measure when setting up a new clinical trial. This will enhance homogeneity and may encourage consensus in the field of IC. RO 18.6 Randomized clinical trials of the impact of alternative diagnostic strategies on patient-important outcomes: a systematic survey El Dib R1 , Tikinnen K2 , Akl E3 , Mustafa RA4 , Agarwal A5 , Gomaa HA6 , Carpenter CR7 , Zhang Y8 , Nascimento Jr P9 , Jorge EC9 , Almeida RA9 , Doles JVP9 , Mustafa AA10 , Sadeghirad B8 , Lopes LC11 , Bergamaschi CC11 , Suzumura EA12 , Cardoso MM9 , Stone SB13 , Schünemann HJ14 , Guyatt GH14 1 Botucatu Medical School, Unesp, Universidade Estadual Paulista, Brazil; 2 Helsinki University Central Hospital and University of Helsinki, Finland; 3 American University of Beirut, Lebanon; 4 University of Missouri-Kansas City, Missouri, USA; 5 University of Toronto, Canada; 6 Tanta University, Egypt; 7 Washington University in St Louis, USA; 8 Department of Clinical Epidemiology and Biostatistics, McMaster University, Hamilton, ON, Canada; 9 Botucatu Medical School - Unesp - Universidade Estadual Paulista, Brazil; 10 Jordan University of Science and Technology, Erbid, Jordan; 11 University of Sorocaba, Sao Paulo, Brazil; 12 Hospital do Coração - HCor, São Paulo, Brazil; 13 Northern Ontario School of Medicine, Canada; 14 Department of Clinical Epidemiology and Biostatistics, McMaster University, Canada Background: Diagnostic tests represent a pivotal part of patient management. Often clinicians adopt tests for clinical Cochrane Database Syst Rev Suppl 1–327 (2015) 104 Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Cochrane Database Syst Rev Suppl 1–327 (2015) 105 use on the basis of findings of diagnostic accuracy. Although in some instances simply knowing accuracy can allow inferences classified as low risk of bias were: generation of allocation (n = 66; 50.7%); allocation concealment (n = 44; 33.8%); blinding (n = 28; 21.5%); missing outcome data (n = 74; 56.9%); selective reporting (n = 44; 33.8%) and; free of other problems (n = 37; 28.4%). Of the 130 RCTs, 44 evaluated mortality; two reported statistically significant results, but neither provided an estimate of relative effect; 28 did not report whether the results were statistically significant and 14 of them reported a non-statistically significant results. Investigators reported the impact of morbidity in 75 RCTs: 16 reported statistically significant results, two of which reported a risk ratio (RR) less than 0.8 and two reported a RR greater than 1.0 –the remaining 14 did not report RRs; 32 did not report whether the results were statistically significant and 27 reported a non-statistically significant results. Conclusions: Randomized trials of diagnostic tests are not rare, seldom show clear benefits on patient-important outcomes, and often suffer from limitations in reporting and conduct. Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. RO 18.7 Social network analysis for identifying central outcomes for clinical research: a case study using Cochrane reviews of HIV/AIDS Saldanha I1 , Li T1 , Yang C2 , Ugarte-Gil C2 , Rutherford G3 , Dickersin K1 1 Cochrane Eyes and Vision Group, USA; 2 Johns Hopkins Bloomberg School of Public Health, USA; 3 University of California, San Francisco, USA Background: The underlying affinity between outcomes that leads to their co-occurrence in a research study identifies central outcomes, i.e. outcomes most important to the connectedness of the network of outcomes. Core outcome sets (COS) are the minimum outcomes that should be measured in research in a field. Current methods to develop COS are inconsistent. Objectives: We conducted a social network analysis (SNA) of outcomes in systematic reviews (SRs) of HIV/AIDS to understand outcome co-occurrence and compare the most central and the most frequent outcomes. Methods: We examined all Cochrane SRs of HIV/AIDS as of June 2013 and grouped individual outcomes Cochrane Database Syst Rev Suppl 1–327 (2015) 106 into 1/14 categories. Our SNA only considered outcomes that co-occurred in ≥ 2 SRs. To identify central outcomes, we used normalized node betweenness centrality (nNBC), i.e. the extent to which connections between other outcomes in a network rely on a given outcome as an intermediary. The higher the nNBC, the more central is the outcome to a network. We identified the 7 most central and most frequent outcomes because the Cochrane Handbook recommends including ≤ 7 main outcomes in SRs. We also examined centrality and frequency of outcomes for the 4 pre-defined HIV/AIDS intervention subgroups: clinical management, biomedical prevention, behavioral prevention, and health services. Results: 140 SRs, measuring 294 unique outcomes, were eligible. The most central outcomes in the overall network were all-cause mortality (nNBC = 23.9) and cost/cost-effectiveness (nNBC = 16.4; Fig 1). The most central and most frequent outcomes differed overall and for each sub-network. For example, for biomedical prevention, adverse events (specified), a patient-important outcome, was among the most central but not among the most frequent outcomes. Only 4/7 outcomes overlapped between the network and frequency analyses for biomedical prevention (Fig 2). Conclusions: SNA is a novel application for systematically identifying central outcomes in SRs. Co-occurrence and frequency are both important considerations for developing COS, particularly because the two contribute different information. SNA should be used when developing COS. RO 18.8 The COMET Initiative database of core outcome sets: increasing its value for users Gargon E1 , Gorst S1 , Williamson P1 1 University of Liverpool, United Kingdom Background: The COMET Initiative developed an online resource to collate the knowledge base for core outcome set (COS) development. It will help authors present their findings clearly and succinctly in Cochrane Reviews, e.g. within the ’Summary of findings’ table. Initial searches (2013) identified 198 COS in different areas of health. Ensuring that the database is as comprehensive as possible and keeping it up to date are key to its value for users. Objectives: To: 1. identify studies published since the original search with the aim of determining which outcomes/domains to measure in all clinical trials in a specific condition; 2. assess website usage in 2014; 3. find out why people are using the database Methods: 1. Searches in MEDLINE, SCOPUS; 2. Analyse website statistics (Google Analytics); 3. A pop-up survey to ask people their reason for searching the database. Results: In January 2015 we searched databases and identified 4980 potentially relevant abstracts. Screening is ongoing to identify the final set of included studies (expected completion April 2015). These results will be compared with the original review to identify any changes in COS development and reporting. Use of the website continues to increase, with more than Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. 16500 visits in 2014 (36% increase over 2013), 12257 unique visitors (47% increase) and 9780 new visitors (43% increase). There has been a rise in the proportion of visits from outside the UK: 8565 visits in 2014 (51% of all visits). By December 2014, a total of 6588 searches had been completed, with 2383 in 2014 alone (11% increase). Pop-up survey is on-going and includes ‘I am planning a systematic review of clinical trials’ as a response option. Conclusions: These updates will bring the database of COS up to date, improving its value for users. It will also highlight clinical areas where gaps exist, providing opportunities for future COS development. We will continue to include studies in an ad-hoc way, until the next update, ensuring that the database is as current as possible. The pop-up survey will provide valuable insight into why people search the database, which will also enable us to improve its value for users. Lunch rapid oral session 5 Primary studies LRO 5.1 Quality of clinical research: a systematic review of definitions and development of a comprehensive framework of indicators von Niederhaeusern B1 , Guyatt G2 , Schandelmaier S3 , Hemkens L3 , Bonde MM1 , Brunner N1 , Rutquist M1 , Bhatnagar N2 , Briel M3 , Pauli-Magnus C1 1 Clinical Trial Unit, Department of Clinical Research, University Hospital, Basel, Switzerland; 2 Department of Clinical Epidemiology and Biostatistics, McMaster University, Canada; 3 Basel Institute for Clinical Epidemiology and Biostatistics, Department of Clinical Research, University Hospital, Basel, Switzerland Background: Clinical research is a prerequisite for advancing our understanding and treatment of disease. While there are factors that some understand as indicators of clinical research quality (e.g. impact factors or whether good clinical practice guidelines were followed), a comprehensive definition considering various stakeholders’ perspectives remains elusive. Objectives: To: (1) systematically review definitions of quality in clinical research; and (2) to create a conceptual framework of indicators allowing for a comprehensive assessment of clinical research quality. Methods: Systematically we searched for definitions of quality in clinical research by screening websites (and any linked information) of various stakeholders involved in clinical research. Stakeholders included governmental bodies, regulatory agencies, pharmaceutical industry, academic and commercial contract research organizations, research ethics committees, patient organizations and funding agencies from 12 countries. In addition, we systematically searched MEDLINE with support from an information specialist. Data synthesis will involve a structured summary of quality definitions. The summary will inform a conceptual framework that we will Cochrane Database Syst Rev Suppl 1–327 (2015) 107 circulate among methodological experts and experienced clinical researchers until no further ideas emerge using the Delphi method. Results: At the time of the Colloquium, we will present how different stakeholders from different countries conceptualize clinical research quality. We will further provide a framework of indicators of clinical research quality that will span all phases of clinical research (i.e. planning to dissemination) and include at least the following dimensions: (1) absence of bias; (2) precision; (3) external validity; (4) innovation/relevance; (5) transparency/public access to data; and (6) well-trained personnel/number of trainees involved. Conclusions: A holistic definition of clinical research quality is still missing. Based on a systematic review we propose a comprehensive framework of quality indicators that may be used as the basis to develop a measurement tool of clinical research quality. LRO 5.2 The time-series studies for health policy research in China 1 1 1 1 Xiuxia L , Jinhui T , Jingchun F , Kehu Y 1 Evidence-Based Medicine Center, School of Basic Medical Sciences, Lanzhou University, China Background: Time series analysis is an effective statistical means for analyzing and forecasting the trends of things, which has been effectively tried and applied in the field of public health. Interrupted time-series studies (ITSs) are usually considered as a valuable research tool for systematic reviews of health systems research. Objectives: Based on the principles of evidence-based medicine to explore and show the evidence of time-series studies in Chinese health policy research. Methods: Four Chinese databases (CBM, CNKI, VIP, WANFANG), four international databases (PubMed, the Cochrane Library, the Campbell Library, Web of Science) and relevant websites were searched in November 2014. The search terms were time-series, health, policy, strategy and China/Chinese. EndNote X4 and Excel were used for data description and analysis. Results: A total of 1232 articles were retrieved; this included 45 (3.7%) time-series studies on national health policies, the first of which was published in 2002, the most recent in 2013 (11; 24.4%), with 35 (77.8%) articles published from 2010 to 2014. Eighteen (40.0%) were concerned with medical insurance, 10 (22.2%) with pharmaceutical policy, seven (15.6%) with hospital management, five (11.5%) wth theoretical exploration, and five (11.1%) with public health. One (2.2%) study was published in the International Journal of Medical Informatics which is included in the Science Citation Index (impact factor 2.716, in 2013), eight (17.8%) were published in the periodicals of the Chinese Science Citation Database, others (80.0%) were published in general periodicals. Conclusions: Our research shows the evidence of interrupted time-series studies in health policy research has an increasing trend year on year in China, but there are fewer concerned with rural health, health reform, health workforce and so on. Very few studies were published Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. in high level periodicals. Further time-series analyses for health policy research should promote and improve this. LRO 5.3 Premature discontinuation of studies approved by research ethics committees: a comparison of randomized and non-randomized clinical studies Blümle A1 , Schandelmaier S2 , Oeller P1 , Kasenda B3 , Briel M2 , von Elm E4 1 German Cochrane Centre, Medical Center, University of Freiburg, Germany; 2 Institute for Clinical Epidemiology and Biostatistics, University Hospital, Basel, Switzerland; 3 Institute for Clinical Epidemiology and Biostatistics, University Hospital, Basel, United Kingdom; 4 Cochrane Switzerland, Institute of Social and Preventive Medicine (IUMSP), University Hospital, Lausanne, Switzerland Background: More than 25% of planned randomized controlled trials (RCTs) are prematurely discontinued. Most discontinuations are not reported to research ethics committees (RECs) and most discontinued RCTs remain unpublished. This raises serious ethical concerns, including waste of scarce resources, loss of collected patient data, and missed opportunities to learn from failure. Similar concerns apply to other prospective studies (non-RCTs) such as non-randomized controlled trials, uncontrolled trials, or cohort studies. However, the prevalence of discontinuation of non-RCTs and the reasons thereof are unknown. Objectives: To assess the prevalence of and the reasons for discontinuation in non-RCTs, and to compare the results to those in RCTs. Methods: Systematically we surveyed studies that were approved by six RECs (Germany, Switzerland, Canada) from 2000 to 2003, enrolled patients or healthy volunteers, and collected outcome data prospectively after study initiation. We had access to study protocols of non-RCTs at one REC (Freiburg, Germany) and to RCT protocols at all six RECs. We will collect study characteristics such as medical area, type of participants, sample size, collaboration, and funding source. We will identify subsequent publications and survey investigators to elucidate whether a study was discontinued and, if so, why. Since some reasons such as benefit or harm only apply to experimental studies, we will stratify the results by study design. Results: An initial screen of study protocols suggests that about 1000 studies were RCTs, 150 uncontrolled trials, 60 non-randomized controlled trials, 25 cohort studies, and 217 ineligible retrospective or cross-sectional studies. We will present data on non-RCTs compared to RCTs regarding the prevalence of discontinuation and the reasons thereof. Conclusions: Compared to RCTs, typical features of non-RCTs likely increase the risk for discontinuation (e.g. first in human study) while others may decrease this risk (e.g. explorative character). Non-RCTs are increasingly included in systematic reviews, so that a a better understanding of their premature discontinuation is warranted. Cochrane Database Syst Rev Suppl 1–327 (2015) 108 LRO 5.4 Are the estimates of blood pressure (BP) lowering effect the same in parallel and cross-over trials? Wong GW1 , Wright JM1 1 Cochrane Hypertension Group, Canada Background: Parallel and cross-over design are two possible designs for randomized control trials assessing the blood pressure (BP) lowering effect of antihypertensive drugs. Each design has strengths and limitations. Both study designs are valid scientific tools to measure the real effect of these interventions. One would expect the effect estimate from parallel studies to be similar to that from cross-over studies. We conducted four systematic reviews on the BP lowering efficacy of beta-blockers. Both parallel and cross-over studies were included in our review so it was possible for us to conduct an indirect comparison of the two types of designs. Method: We compared the BP lowering efficacy from parallel and cross-over studies within the same subclass of beta-blockers. The data were synthesized using RevMan 5.2 and the mean estimates from the two designs were compared using Student’s t tests. Results: The analysis included 97 trials that examined the BP lowering efficacy of 20 different beta-blockers. To be included the studies had to be randomized, double-blind, placebo-controlled trials in primary hypertensive patients taking fixed dose beta-blockers for 3 to 12 weeks. The BP lowering estimates (systolic/diastolic) from cross-over trials exceeded those from parallel trials for all subclasses: nonselective beta-blockers -12/-9 mmHg vs -5/-4 mmHg, alpha and beta receptor blockers -11/-7 mmHg vs -4/-3 mmHg, partial agonists -9/-5 mmHg vs -7/-3 mmHg and beta-1 selective blockers, -13/-11 mmHg vs -7/-6 mmHg. The overall weighted mean was -13/-9 mmHg and -7/-5 mmHg respectively. All the comparisons were statistically significant except for the partial agonists. Conclusion: The findings suggest that the BP lowering effect of beta-blockers is greater by 6/4 mmHg in cross-over trials than parallel trials. This cannot be a real effect and creates a serious dilemma for reviewers who are pooling results from cross-over and parallel trials. In this case we suspect that the estimate from the parallel trials is closer to the real effect. These findings need to be replicated in other reviews. It is possible that results from cross-over trials are not reliable. LRO 5.5 Are routinely collected data studies reliable surrogates when there are no randomized trials? A meta-epidemiological study Hemkens LG1 , Contopoulos-Ioannidis DG2 , Ioannidis JPA1 1 Stanford Prevention Research Center, Department of Medicine, Stanford University School of Medicine, Stanford, California, USA; 2 Department of Pediatrics, Division of Infectious Diseases, Stanford University School of Medicine, Stanford, California, USA Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Background: Routinely collected data (RCD) studies are proposed to inform health care decisions when randomized controlled trials (RCTs) are not timely available, but have an inherent risk of selection bias due to confounding by indication. Propensity-score-analyses are frequently used to address selection bias issues, but biases due to remaining unaccounted confounders cannot be excluded. Objectives: To assess the agreement between results of RCD-studies and subsequent RCTs on the same topic in areas without RCT evidence to guide health care decisions. Methods: We searched PubMed (up to November 2014) for RCD-studies published up to 2010 that used propensity scores and reported comparative effects of medical interventions for mortality endpoints. We included RCD-studies that were conducted before any RCT was published for the same topic and same compared interventions. We searched systematically for subsequently published pertinent RCTs, extracted mortality data, and calculated the mortality odds ratio (OR). When more than one subsequent RCT was identified, we combined them with random-effects models (or Peto’s approach for rare events) and calculated the summary OR. We analyzed the agreement between RCD-studies and RCTs by calculating the relative OR (ROR; i.e. summary OR of RCT(s) divided by RCD-study estimate; ROR > 1 means RCD overestimate survival benefits). We then synthesized the individual ROR data across all RCD-RCT pairs to calculate the summary ROR (sROR) as overarching measure of the agreement of early RCD-effects and subsequent RCT-results. Results: We identified 929 records, evaluated 420 in full-text, and screened 124 RCD-studies further. We analyzed 16 RCD-studies without preceding RCTs and for which subsequent pertinent RCTs were identified. Preliminary results indicate limited agreement between RCD-effects and subsequent RCT evidence. RCD-studies showed significantly inflated results (sROR 1.31; 95% CI 1.04 to 1.65). Conclusions: RCD-studies that have been conducted prior to RCTs seem to systematically and substantially overestimate the benefits of medical treatments. Final results will be available at the Cochrane Colloquium 2015. LRO 5.6 Quasi-randomisation in emergency setting trials: a recipe for selection bias, or an efficient approach? Corbett M1 , Moe-Byrne T1 , Oddie S1 , McGuire B1 1 Centre for Reviews & Dissemination, University of York, United Kingdom Background: Recruitment into randomised trials in emergency settings may be difficult because of the inherent time restrictions encountered when enrolling, randomising and treating participants. Although quasi-randomisation might mitigate some of these problems, systematic reviewers are likely to judge such trials as having a high risk of bias, due to the possibility of selection bias. However, selection bias may not actually occur since the very nature Cochrane Database Syst Rev Suppl 1–327 (2015) 109 of emergency settings might preclude opportunities for bias. Objectives: To compare quasi-randomisation with random sequence generation when used in emergency setting trials, on the prevalence of possible selection bias, and on trial recruitment outcomes. Methods: Systematic reviews of participants with an acute injury or illness, requiring intervention as quickly as was clinically practicable were identified using the ‘emergency medicine’ topic in the Cochrane Library. Eligible reviews incorporated at least one trial using truly random sequence generation and at least one trial using quasi-randomisation. Evidence of possible selection bias was identified by assessment of group baseline characteristics. Clinical advice and published studies were used to identify important prognostic indicators, and important magnitudes of group difference. Results: The seven eligible reviews included 27 eligible trials: 11 used quasi-randomisation and 16 used random sequence generation. Important group imbalance was identified in two of the 11 quasi-randomised trials (18%) and four of the 16 trials using true sequence generation (25%); of the latter, three trials described appropriate methods to conceal treatment allocation, though all three also had small sample sizes. Clinical heterogeneity and poor reporting limited the assessment of trial recruitment outcomes. Conclusions: In emergency care settings quasi-randomised trials may be no more likely to result in biased recruitment of participants than trials using true randomisation. The likelihood of chance imbalances affecting results is arguably more important to consider, for both systematic reviewers and trial investigators. Lunch rapid oral session 6 Health Economics LRO 6.1 Landscape review of costing methods for reproductive, maternal and child health interventions Das J1 , Bhutta Z1 1 Aga Khan University, Pakistan Background: Cost information is essential to improve the economic efficiency of healthcare systems and costing studies play a major role in this regard. However the methods used across different economic evaluation studies are diverse, and it is it is imperative to understand these methods and their applicability in different settings. Objectives: We undertook a review to evaluate various cost scale-up evaluations from regional, national and multinational scale-up studies for Reproductive, Maternal, Newborn and Child Health (RMNCH) interventions to identify the major methods used for estimating the costs. We also aimed to compare the features of each costing method using an evaluation matrix based on peer review and publication quality. Methods: Peer-reviewed literature search was conducted for studies on scale-up costing for RMNCH. A Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. data abstraction table was created to extract details on the objective and components of the method, as well as the areas of application for each method. Scale-up methods were then categorized and tabulated. Results: We included 24 studies focusing on regional, national or multi-country scale-up of RMNCH interventions. We prepared a summary of the components of the evaluation matrix from each of these studies. Cost scale-up projections for RMNCH interventions are difficult, however various methods have been used to estimate the scale-up costs for evidence-based decision making. Different cost concepts and costing methodology are being used depending upon the purpose for which cost data are being used, the perspective of the study, the type and complexity of the health service, the precision required, the requirements of generalizability and representativeness and the availability of reliable and valid data. Conclusions: Large scale regional and global cost scale-up analysis mostly relies on scenarios of how results vary under certain situations or on averaging the results from a set of models. There is a need for a standardized costing methodology, if all stakeholders, including providers, purchasers and policy makers, are to make informed decisions. Accurate costing can contribute to the efficient allocation of resources. LRO 6.2 HPV vaccines in low- and middle-income countries: a synthesis study of economic evaluations Canelo-Aybar C1 , Mezones-Holguin E1 , Balbin G2 , Perez-Gomez A3 , Florez ID4 1 School of Medicine, Universidad Peruana de Ciencias Aplicadas, Peru; 2 Casimiro Ulloa, Peru; 3 Health Technology Assessment Institute, Colombia; 4 McMaster University, Canada Background: HPV vaccines are an important intervention for preventing cervical cancer. In public health, associated costs should be included for making decisions. Thus, economic evaluations (EE) play a relevant role, especially in the context of low- and middle-income countries (LMICs). Objectives: To describe and appraise critically the available published EE evidence on HPV vaccines in LMICs. Methods: A systematic literature review using a validated search strategy was carried out to identify a complete EE (cost-effectiveness, cost-utility and cost-benefit analysis) for HPV vaccines in LMICs published up to March 2015 in PubMed/MEDLINE, EMBASE and the Cochrane Library. Language restrictions were not used. Congress abstracts were excluded. Quality assessment was performed using the Consensus on Health Economic Criteria instrument. Also, age at which vaccinated, analysis perspective, clinical and economical outcomes, type of vaccine, model performed, Incremental cost-effectiveness ratio (ICER), and founding were analyzed, as well as other variables. Results: A total of 1827 citations were obtained, of which 415 were duplicates. After title and abstract screening, 64 unique studies were reviewed (Fig 1) and 43 EEs were included for full qualitative synthesis -18 additional studies Cochrane Database Syst Rev Suppl 1–327 (2015) 110 from a previous systematic review. Only seven studies were done specifically in low-income countries. Vaccination age ranged between from 10 to 15 years old. Most investigations incorporated only cervical cancer (32), but others additionally included harms (11) in their analysis. Most studies (34) were funded by public resources and indirect costs were assessed in 23 studies. The static models are the most frequent technique used and 10 studies had a multi-country evaluation. In most studies, findings suggested that vaccination is likely to be cost-effective; however these results are dependent on context and can be affected by GDP per head and existing cervical cancer screening program. Conclusions: Beyond the significant heterogeneity and limited quality of current evidence, there is a tendency to support the cost-effectiveness of HPV vaccines in these countries. LRO 6.3 Pharmaceutical policies: effects of reference pricing, other pricing, and purchasing policies (Updated Review) Acosta A1 , Ciapponi A2 , Aaserud M3 , Vietto V4 , Austvoll-Dahlgren A5 , Kösters JP6 , Vaca C7 , Machado M8 , Diaz Ayala DH8 , Oxman A5 1 School of Pharmacy, Universidad Nacional de Colombia, Colombia; 2 Argentine Cochrane Centre IECS, Institute for Clinical Effectiveness and Health Policy, Hospital Italiano de Buenos Aires, Southern American Branch of the Iberoamerican Cochrane Centre, Buenos Aires, Argentina; 3 Statens Legemiddelverk, Norwegian Medicines Agency, Oslo, Norway; 4 Division of Family and Community Medicine, Hospital Italiano de Buenos Aires, Argentina; 5 Norwegian Knowledge Centre for the Health Services, Oslo, Norway; 6 The Nordic Cochrane Centre, Rigshospitalet, Copenhagen, Denmark; 7 Grupo RAM, Universidad Nacional de Colombia, Bogota, Colombia; 8 Departamento de Farmacia, Universidad Nacional de Colombia, Bogota, Colombia Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Background: Pharmaceuticals are important interventions that could improve people’s health. Pharmaceutical pricing and purchasing policies are used as cost-containment measures to determine or affect the prices that are paid for drugs. Internal reference pricing establishes a benchmark or reference price within a country which is the maximum level of reimbursement for a group of drugs. Other policies include price controls, maximum prices, index pricing, price negotiations and volume-based pricing. Objectives: To determine the effects of pharmaceutical pricing and purchasing policies on health outcomes, healthcare utilisation, drug expenditures and drug use. Methods: Policies in this review were defined as laws; rules; financial and administrative orders made by governments, non-government organ- isations or private insurers. To be included a study had to include an objective measure of at least one of the following outcomes: drug use, healthcare utilisation and health outcomes or costs (expenditures); the study had to be a randomised trial, non-randomised trial, interrupted time series (ITS), repeated measures (RM) study or a controlled before-after study of a pharmaceutical pricing or purchasing policy for a large jurisdiction or system of care. Results: We included 18 studies (seven identified in the update): 17 of reference pricing, one of which also assessed maximum prices, and one of index pricing. None of the studies were trials. All included studies used ITS or RM analyses. The quality of the evidence was low or very low for all outcomes. Four studies reported effects on mortality and healthcare utilisation, however they were excluded because of study design limitations. Conclusions: The majority of the studies of pricing and purchasing policies that met our inclusion criteria evaluated reference pricing. We found that internal reference pricing may reduce expenditures in the short term by shifting drug use from cost share drugs to reference drugs. Index pricing may reduce the use of brand drugs, increase the use of generic drugs, and may also slightly reduce the price of the generic drug when compared with no intervention. Reference Acosta A, Ciapponi A, Aaserud M, Vietto V, Austvoll-Dahlgren A, Kösters JP, Vacca C, Machado M, Diaz Ayala DH, Oxman AD. (2014) Pharmaceutical policies: effects of reference pricing, other pricing, and purchasing policies (Review). The Cochrane Library, Issue 10. LRO 6.4 Calculating costs: facing challenges for a cost comparison of day surgical and inpatient varicose vein surgery as an example Fischer S1 1 Ludwig Boltzmann Institute for Health Technology Assessment, Austria Background: Evidence-based medicine (EbM) and Health Technology Assessments (HTA) aim to improve health care and allocation of resources, so besides clinical aspects, Cochrane Database Syst Rev Suppl 1–327 (2015) 111 cost calculations are becoming more important, with selection of data being crucial for an appropriate calculation. Objectives: Comparison of the costs of day surgical and inpatient varicose vein surgery as an example to show the results of using different data sources. Methods: For the first approach, we used meta-data from the Austrian ‘‘Dokumentations- und Informationssystem für Analysen im Gesundheitswesen’’. We considered generated lump sums from hospital reimbursement, total costs that occurred in the departments and number of patients in an equation. For the second approach, we used data from individual hospitals to calculate the costs for the surgery itself and nursing. The third approach contained the adaption of international cost data, by an adjustment for inflation and prices for Austria. Results: The calculated costs of varicose vein surgery differed between EUR 859 and EUR 4664 for a day case and EUR 1720 to 2330 for an inpatient treatment –depending on the used approach. The main strength of the first approach is that it can be done relatively quickly. However, the validity of the calculated costs is low. Thus, a difference between real and calculated costs is likely. The use of hospital data takes more time, though, the quality of the data is much better. A weakness of this approach is that individual costs are not available for a specific intervention. Furthermore, these costs are hospital specific and generalising for other hospitals is difficult. The fast acquisition of the international reference costs is a main strength, though, the costs are from a different healthcare system and therefore, the transferability of the costs is limited. Conclusions: The results have shown that an examination of administrative data is indispensable for proper cost evidence. Calculations of costs for EbM and HTA that are adjusted to national circumstances are required. However, many issues of an efficient healthcare system and appropriate allocation of resources are unsolved. LRO 6.5 Impact of pay-for-performance on behavior of primary care physicians and patient outcomes: a systematic review Liao G1 , Lin Y2 , Huang J3 , Lin Y2 1 West China Hospital of Stomatology, Sichuan University, China; 2 Department of Urology, West China Hospital, Sichuan Univerisity, China; 3 West China Hospital, Sichuan Univerisity, China Background: Pilot countries such as the UK have suggested that pay-for-performance (P4P) is an effective payment method, which links physicians’ income to the quality of their services. With its wide application in primary health care, more countries are exploring its role among general practitioners. Interestingly, research on P4P has yet to reach an agreement, so a systematic review was performed for a more affirmative result. Objectives: To provide a comprehensive and objective evaluation of P4P for decision-makers. Methods: Studies were identified by searching PubMed, EMBASE and the Cochrane Library. Electronic searching was conducted in the fourth week of January 2013. As the included studies had significant Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. clinical heterogeneity, a descriptive analysis was conducted, and a Quality Index was used to assess the quality of the evidence. Results: Among the 44 included articles, 36 were about the effect of P4P on quality of chronic diseases; 20 were about equity, five about satisfaction of patients, four about cost-effectiveness, 24 about a factor of implementation of P4P(the baseline of P4P and the practice size). P4P had an overall positive effect, which varied in accordance with the baseline quality of medical care and the practice size. The implementation of P4P could bring about new problems regarding the inequity of medical care, patients’ dissatisfaction and the rise of medical cost. Conclusions: Decision makers should consider the baseline conditions of medical quality and the practice size before new medical policies are enacted. Furthermore, most studies investigated here were retrospective and observational with high levels of heterogeneity, though the descriptive analysis is still of significance. LRO 6.6 Network meta-analysis of multiple outcomes to inform a cost-effectiveness analysis: interventions for the induction of labour Keeney E1 , Alfirevic Z2 , Dowswell T2 , Welton NJ1 , Medley N2 , Dias S1 , Jones LV2 , Gyte G2 , Caldwell DM1 1 University of Bristol, United Kingdom; 2 University of Liverpool and Liverpool Women’s Hospital, United Kingdom Background: Network meta-analysis (NMA) allows estimation of multiple intervention effects across a network of interventions. NMA is particularly useful for economic evaluations, which compare the cost-effectiveness of multiple interventions. However, economic evaluations usually depend on several outcomes, whereas a systematic review typically reports one outcome at a time. Objectives: Our aim is to illustrate the joint modelling of multiple outcomes, and assessment of inconsistency in NMA, for providing inputs to an economic model. Methods: We use as an example a systematic review and NMA to identify the most cost-effective intervention(s) for induction of labour. Trials in this area report multiple, but related, outcomes for the mother and baby. The economic model requires estimates of the probabilities of vaginal delivery (VD) within 24 hours, VD after 24 hours, and caesarean section, as well as the proportion of babies admitted to intensive care. These outcomes are not independent, and require estimating jointly in the NMA. Careful definition of denominators allows estimates of conditional probabilities for inputs for the economic model to be obtained. We use various approaches to explore and deal with inconsistency in the NMA, including node-splitting, continuity correction and removing trials at high risk of bias. Results: Titrated (low-dose) oral misoprostol solution is the treatment that not only has the lowest cost but also results in the highest utility for mothers and babies. Conclusions: Related outcomes should be jointly modelled in order to provide the required estimates for economic models. We Cochrane Database Syst Rev Suppl 1–327 (2015) 112 found that trials did not always provide enough information to be able to identify the outcomes needed in the economic model. Exploring and dealing with inconsistency is essential for the validity of conclusions drawn from an economic evaluation based on a NMA. LRO 7.2 Policymakers’ familiarity with and preference for overall quality of evidence definitions Kriz HM1 , Harrod CS1 , Gerrity MS1 , King VJ1 1 Oregon Health & Science University, USA Lunch rapid oral session 7 Involvement of users and stakeholders LRO 7.1 Patient Involvement in systematic reviews –why it is so important? Rhodes C1 , Jordan JL1 , Belcher J1 , Yardley S1 , Hyde C1 , Higginbottom A1 , Worrall A1 , Taylor R1 1 Research Institute for Primary Care and Health Sciences Keele University, United Kingdom Background: The Primary Care and Health Sciences Research Institute at Keele University in the UK has an established patient and public involvement (PPI) group informing research studies. Objectives: To describe how patients developed skills to understand research methods patients were involved in two systematic reviews: 1. narrative synthesis: shared decision-making in primary care consultations; and 2. realist synthesis: learning and healthcare delivery in primary care. Methods: Support for patients included workshops on systematic reviews, evidence-based medicine and understanding statistics. A PPI coordinator assessed needs throughout, ensuring adequate time and support within project timescales. Narrative synthesis: five patients participated in three sequential workshops ensuring the research question was relevant and that factors important to patients in sharing decisions were identified. Realist synthesis: two patients, a PPI co-ordinator and researcher produced a data extraction sheet useable by all team members and papers were split between them to review. Results: Narrative synthesis: patients identified additional factors, poorly represented in the literature and planned dissemination of results with different audiences. Realist synthesis: patients ensured data extraction forms were developed in an easy to read format and formed a seminar panel to disseminate methods used, informing future primary care delivery on multimorbidity. Conclusions: Patients working within a research team can inform the review question and undertake extraction, interpretation and analysis of data. Patients involved in research need support to enable them to develop skills to make sense of scientific evidence. Patients bring a unique perspective to many areas of the research review process, identifying areas of importance. Researchers need support to translate research terms into accessible formats to ensure the patient voice is truly integrated into their review. Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Background: Oregon Health & Science University’s Center for Evidence-based Policy (Center) conducts evidence summaries for policy makers using the Grading of Recommendations Assessment, Development and Evaluation (GRADE) approach to assess overall quality of evidence (QoE). Many policymakers find it difficult to interpret and apply GRADE QoE definitions. Objectives: Assess policymakers’ familiarity with GRADE and preference for five different QoE definitions. Methods: Policymakers participating in Center projects were invited to complete a 24-item web-based survey assessing familiarity and preference of five QoE definitions including: GRADE, Agency for Healthcare Research and Quality (AHRQ), Evidence-based Practice Center (EPC), and three alternative definitions developed by the Center. Familiarity with GRADE was based on a 2-item composite variable. Policymaker degree of certainty that findings will remain stable over time was rated from 0% to 100% for each QoE definition. Odds ratios were used to assess the association between familiarity and definition preference. Results: Out of 92 policymakers, 42 (46%) completed the survey with 25 (60%) indicating that they were familiar with GRADE. Nearly half of respondents preferred the EPC definitions. Fifteen (38%) preferred alternative definitions, but none selected a brief lay terminology definition. Only five (13%) preferred the standard GRADE definition. For each GRADE QoE level, the median per cent certainty that findings from evidence would remain stable was 90% for high QoE, 75% for moderate QoE, and 45% for low QoE. Although 29 (69%) policymakers had experience with GRADE, none were extremely familiar with it. Familiarity with GRADE was not associated with a preference for standard (GRADE or EPC) QoE definitions compared to alternatives (OR = 1.3, 95% CI 0.4 to 5.1; P value 0.7). Conclusions: The EPC definitions were preferred by policymakers. Despite having experience with GRADE, no policymaker felt extremely familiar with it. Policymakers’ familiarity was not associated with definition preference. Strategies to enhance policymakers’ understanding and application of GRADE ratings are needed. LRO 7.3 An approach for systematic review teams to engage policymakers and stakeholders: from question selection to knowledge translation El-Jardali F1 , Fadlallah R1 , Akl E1 1 American University of Beirut, Lebanon Background: Engaging policymakers and stakeholders in evidence synthesis activities is an essential step to Cochrane Database Syst Rev Suppl 1–327 (2015) 113 promote the use of evidence in policy making and practice. This creates a need for a well-developed and structured approach that spans from selecting relevant review questions and synthesizing evidence, to using it in informing policy decisions. Objectives: To describe an innovative approach to engage policy makers and stakeholders in evidence synthesis activities and knowledge translation. Methods: The development methodology followed a multifaceted process that encompassed the following steps: 1. comprehensive literature review; 2. development of approach; 3. expert opinions; 4. refinement of approach; and 5. real life experience (implementation). Results: The proposed approach is shown in Figure 1. It is broadly divided into three phases: 1. preparatory work for selection of potential review questions; 2. priority setting and evidence synthesis; and 3. knowledge translation. The approach is characterized by its comprehensiveness, the ongoing engagement of Figure 1 policymakers and stakeholders, and the segregation of review questions into those for systematic reviews and those for knowledge translation. In the presentation, we will provide a detailed description of the approach, including the process and outcome of its development. In addition, we will draw on case studies to demonstrate the application of the approach. We will also discuss lessons learned and implications for other groups and systematic review centers planning to engage policymakers and stakeholders in evidence synthesis activities. Conclusions: The proposed approach will help align the production of evidence syntheses with policy Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. priorities and promote their use in decision-making. It can also help leverage and harness existing systematic reviews on priority questions and devise appropriate knowledge translation products. LRO 7.4 The challenges of using evidence-based methods for decisions at a regional level involving stakeholders without research training Grillich L1 , Sommer I2 , Mahlknecht P1 , Gartlehner G1 1 Danube University, Krems, Austria; 2 Danube University, Krems, Austria Background: Childhood obesity is a main public health challenge. Children who are overweight or obese are at greater risk of poor health in adolescence and also in adulthood. Thus, the regional government of Lower Austria wanted to implement an evidence-based decision process involving stakeholders and experts whether or not to implement a life-style intervention program for obese children/adolescents. Objectives: To describe the challenges of implementing an evidence-based decision process using GRADE (Grading of Recommendations Assessment, Development and Evaluation) and DECIDE (Developing and Evaluating Communication Strategies to Support Informed Decisions and Practice Based on Evidence) frameworks at a regional level involving local stakeholders without prior knowledge of evidence-based or other research methods. Methods: We used GRADE and DECIDE to appraise and communicate the quality of evidence and the strength of recommendations. To assess the relative importance of the desirable and undesirable outcomes we conducted two focus groups (one with overweight children/adolescents and one with parents of these). Scientific experts and relevant stakeholders selected and prioritized outcomes. We carried out an umbrella review to critically appraise, synthesise and grade the current evidence and guided a shared evidence-based recommendation making process with all stakeholders. Results: Using GRADE for a complex public health question at a regional level was challenging: we had to convey to experts and stakeholders that prioritizing outcomes is necessary for decision making and we had to disentangle the large number of prioritized outcomes and map them to potentially quantifiable outcomes. Last but not least, the heterogeneity of outcome measures reported in the literature required the adaptation of the GRADE system to rating the quality of evidence. Conclusions: Although using the GRADE and DECIDE frameworks at a regional level involving stakeholders without research training was challenging, it provided a structured and well received method to reach a conclusion about implementation of a life-style intervention program for obese children/adolescents. Cochrane Database Syst Rev Suppl 1–327 (2015) 114 LRO 7.5 Extensive consumer/stakeholder engagement leads to practical key messages Van Eerd D1 , Irvin E2 , Munhall C2 , King T1 1 Institute for Work & Health, Canada; 2 Cochrane Back Review Group, Canada Background: The goal of systematic reviews is to provide an up-to-date synthesis of the evidence that can be used in practice to improve the health of people. Individuals who have a stake in the evidence include researchers, providers/ clinicians and consumers/clients. Changing practice involves consumers/clients, therefore providing information in a way that is useful to these stakeholders is paramount. Objectives: Describe the engagement of consumers/stakeholders in a review process and the resulting changes in key messages. Methods: This is a case study using a review of the effectiveness of workplace prevention of upper extremity musculoskeletal disorders employing a Cochrane approach. In this review, we adapted a stakeholder engagement process (Keown 2008). The engagement process was iterative and engaged consumers/stakeholders in the co-development of key messages from the review. Results: An iterative approach was used to engage multiple consumers/stakeholders at multiple time-points through the review process. In particular, consumers/stakeholders were heavily engaged in the generation of relevant and practical messages that they could implement in their day-to-day practice. They were also instrumental in determining and accepting the threshold for sufficient evidence and accompanying messages. Conclusions: An iterative approach resulted in practical messages and recommendations. Through this iterative process messages became more practical and ‘useful’ to consumers/stakeholders. While engaging with multiple consumers/stakeholders is time-consuming we experienced greater uptake of the review results and key messages with this process. ensure they are appropriate and accessible to different population groups, but the involvement of people with the condition is not always considered when designing the intervention. Appropriateness, relevance and acceptability of interventions are not routinely considered as criteria when conducting reviews, but may contribute to variations in effectiveness. Objectives: Our realist review aimed to develop a preliminary theory of the contexts that promote/enable patient involvement in diabetes research, and whether involvement could produce more relevant interventions. Our review questions explored how people have been involved in diabetes research, and whether involvement influenced the development and implementation of diabetes interventions. Methods: We assembled a review team of diabetes researchers who are seeking information on involvement to inform the design and conduct of their studies. A preliminary theory for successful involvement was developed from involvement literature. We searched MEDLINE, CINAHL, EMBASE in the first instance to identify articles discussing involvement with people at risk for/with diabetes. Cluster searching identified multiple articles on the same study. Several typologies for patient involvement were tested to determine their utility as data extraction tools. Results so far: To date we have included 79 articles. Cluster searching identified eight projects that generated multiple articles (37). Two different approaches to involvement are used: a community-based participatory approach where people are actively involved in all stages of development and implementation, and a more ‘traditional’ approach where researchers selectively involve people at specific points. Outcomes chaining is being used to identify relationships between the involvement process, uptake and attainment of diabetes-related goals. Lunch rapid oral session 8 Search filters LRO 8.1 LRO 7.6 Search filters: mapping the literature using a novel form of content analysis Does involvement in knowledge creation contribute to development of complex interventions? A realist review of involvement in diabetes research Wilson J1 , Anderson M1 1 CDPLPG, United Kingdom Harris J1 , Graue M2 , Kirkevold M3 , Haltbakk J2 , Rokne B4 , Dunning T5 1 Qualitative&ImplementationResearchMethods, UnitedKingdom; 2 Bergen University College, Norway; 3 University of Oslo, Norway; 4 University of Bergen, Norway; 5 Deakin University, Australia Background: Systematic reviews are considered as the foundation of knowledge translation, but their quality can only be as good as the primary studies that are included in the review. Recently there have been several reviews looking at the importance of tailoring interventions to Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Background: Search filters are combinations of words and phrases designed to retrieve an optimal set of records on a particular topic (subject filters) or study design (methodological filters). Information specialists are increasingly turning to reusable filters to focus their searches. However, the extent of the academic literature on search filters is unknown. We provide a broad overview to the academic literature on search filters. Objectives: To map the academic literature on search filters from 2004 to 2015 using a novel form of content analysis. Methods: We conducted a comprehensive search for literature between 2004 and 2015 across eight databases using a subjectively derived search strategy. We Cochrane Database Syst Rev Suppl 1–327 (2015) 115 identified key words from titles, grouped them into categories, and examined their frequency and co-occurrences. Results: The majority of records were housed in Embase (n = 178) and MEDLINE (n = 154). Over the last decade, both databases appeared to exhibit a bimodal distribution with the number of publications on search filters rising until 2006, before dipping in 2007, and steadily increasing until 2012. Few articles appeared in social science databases over the same time frame (e.g. Social Services Abstracts, n = 3). Unsurprisingly, the term ‘search’ appeared in most titles, and quite often, was used as a noun adjunct for the word ‘filter’ and ‘strategy’. Across the papers, the purpose of searches as a means of ‘identifying’ information and gathering ‘evidence’ from ‘databases’ emerged quite strongly. Other terms relating to the methodological assessment of search filters, such as precision and validation, also appeared albeit less frequently. Conclusions: Our findings show surprising commonality across the papers with regard to the literature on search filters. Much of the literature seems to be focused on developing search filters to identify and retrieve information, as opposed to testing or validating such filters. Furthermore, the literature is mostly housed in health-related databases, namely MEDLINE, CINAHL, and Embase, implying that it is medically driven. Relatively few papers focus on the use of search filters in the social sciences. LRO 8.2 UK filter project: methodological approach to reduce information overload by developing validated UK geographic search filters for MEDLINE and Embase Ayiku L1 , Craven J1 , Levay P1 , Finnegan A1 , Barrett E1 1 National Institute for Health and Care Excellence (NICE), United Kingdom Background: Studies from United Kingdom (UK) health and social care settings can be required to inform evidence-based guidance developed by the UK National Institute for Health and Care Excellence (NICE). However, robust methods for retrieving UK studies in systematic literature searches do not exist. The pragmatic use of non-validated UK search strategies risks an unknown proportion of UK studies being missed. Conversely, sifting search results that are not restricted by geographic location creates a costly and time-consuming information overload when UK studies are required. Information specialists at NICE are developing validated UK geographic search filters to resolve this issue. Objectives: To develop validated UK geographic search filters for MEDLINE and Embase (Ovid) with the best balance of recall and precision for systematic literature searches. Methods: 1. Gold standard: This will be generated using the relative recall method. UK studies that have informed NICE guidance will be identified and the studies that are indexed in MEDLINE or Embase will form the gold standard for each database. A power calculation will determine the number of gold standard studies required to develop each filter. 2. Filter development: The gold standard for each database will be randomly divided Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. into a ‘development set’ and a ‘test set’. The UK-related search terms for the MEDLINE and Embase filters will be derived objectively using a text analysis system on the ‘development set’ of each database. 3. Validation: The effectiveness of the filters will be validated against the ‘test set’ of each database by calculating their recall and precision. Expected conclusions: The validated UK geographic search filters will enable the robust retrieval of UK studies from MEDLINE and Embase in systematic literature searches. The filters and accompanying methods will be shared via online search filter resources and peer-reviewed journal articles. Professional peers will be able to use the filters to retrieve UK studies or to transfer the development methods to create search filters for other geographic regions. LRO 8.3 Feasibility of developing and validating a geographic filter for a small country like Austria Jeitler K1 , Semlitsch T2 , Horvath K3 , Posch N2 , Berghold A4 , Siebenhofer A2 1 Institute of General Practice and Evidence-Based Health Services Research and Institute for Medical Informatics, Statistics and Documentation, Medical University of Graz, Austria; 2 Institute of General Practice and Evidence-Based Health Services Research, Medical University of Graz, Austria; 3 Institute of General Practice and Evidence-Based Health Services Research and Department of Internal Medicine, Division of Endocrinology and Metabolism, Medical University of Graz, Austria; 4 Institute for Medical Informatics, Statistics and Documentation, Medical University of Graz, Austria Background: In search strategies focusing on studies conducted within local health care systems highly sensitive geographic filters (GFs) may be a reasonable means to reduce research resources. At present, only few validated GFs are available and to our knowledge no GF for Austrian studies has been published. Objectives: To develop and validate a GF to retrieve studies with Austria as a reference point for use in the PubMed database, if feasible. Methods: At first, we will estimate roughly the prevalence of Austrian publications among all PubMed entries in the last five years searching for the term ’Austria’ in the affiliation [ad], title and abstract [tiab] fields as well as in the geographic locations category of the medical subject headings [MeSH]. We will consider further steps to be appropriate, if at least 10 relevant publications in a validation reference set of 2000 abstracts can be expected. We will limit the reference set to a workable number of 2000 abstracts as the abstracts have to be categorized manually by relevance on the basis of the information available from the database. The GF to identify studies performed in Austria will be created using English and German terms derived from the names of the country, its provincial states, and districts. These terms will be searched for in additional fields like transliterated title [tt] or other term [ot] besides the aforementioned fields. A random sample of 2000 references will be drawn from articles published in PubMed within the last five years. Two reviewers will independently identify Cochrane Database Syst Rev Suppl 1–327 (2015) 116 relevant articles for the reference set. We then will assess the performance of the Austrian GF in terms of its specificity and sensitivity. Results: We will present results on whether developing a geographic filter for a small country like Austria is feasible with an acceptable amount of effort and discuss possible caveats. Furthermore, filter details and performance data will be published. LRO 8.4 Finding qualitative literature on dementia in MEDLINE, EMBASE, PsycINFO and CINAHL: assessment of three qualitative search filters Rogers M1 , Bethel A1 University of Exeter Medical School, United Kingdom 1 Background: Systematic reviews of qualitative data and mixed methods reviews play an important role in dementia care and care planning. Finding qualitative data in medical databases to inform reviews can be difficult. Objectives: 1. To investigate whether qualitative search filters help with finding qualitative research in dementia. 2. To find out which filter works best in terms of effectiveness and ease of use. 3. To examine the effectiveness of three qualitative filters on four major databases (MEDLINE, EMBASE, PsycINFO and CINAHL). Methods: We identified qualitative or mixed methods systematic reviews in dementia care by searching the Cochrane Database of Systematic Reviews (CDSR), the Database of Abstracts of Reviews of Effects (DARE), the Health Evidence database (http://www.healthevidence.org/), MEDLINE, EMBASE, PsycINFO and CINAHL. Qualitative studies identified in these reviews were used to form a test set of literature. Three qualitative methodology filters were selected (one with few, broad-based terms, one with many specific free-text terms and thesaurus terms for qualitative studies), adapted for the four databases, and were tested for sensitivity, precision and specificity against the test literature. Results: The results for the performance of each filter for each database will be presented. Conclusions: This work will help to inform whether it is useful to utilise a search filter to identify qualitative literature in dementia, and which is the most effective filter for locating this literature in four major databases. LRO 8.5 Use of filters in electronic searching for observational studies Jaiswal N1 , Singh M1 , Thumburu K2 , Kaur J2 , Chadha N3 , Kumar M2 , Agarwal A2 1 Department of Pediatrics & ICMR Advanced Centre for Evidence-Based Child Health, Post Graduate Institute of Medical Education and Research, India; 2 ICMR Advanced Centre for Evidence-Based Child Health, Post Graduate Institute of Medical Education and Research, India; 3 Dr.Tulsidas Library, Post Graduate Institute of Medical Education and Research, India Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Background: Searching various databases is an exhausting job for researchers. Different databases require different terminologies and different filters and multiple databases need to be searched for ensuring comprehension and avoiding bias. The search strategies formulated are therefore long and involve great expertise. Using search filters provided by the databases is therefore an aid to decrease the load. There are available search terms for searching for randomized controlled trials, but no such standardized strategy exists for searching for observational studies. Objectives: Comparison of search outputs with or without using the search filters in searches involving observational studies. Methods: We illustrate the use of search filters and searching without the use of specific filters through an example of a systematic review of observational studies. The databases searched were Pubmed, Embase and MEDLINE via Ovid SP. Two reviewers searched the databases independently, one used the search filters provided by the database and the second reviewer did not use the specific filters, but instead used the specific search terms for the observational studies for electronic searching of the databases. The other blinded reviewers then compared and screened the results. Results: Use of specific search filters significantly decreased the total number of hits from 16323 records to 6423 records and did away with most of the irrelevant results, but missed out on 5% of the included studies. Conclusions: The specific search filters reduce the workload, but need to be used with caution as some of the important records may be missing. Acknowledgements: This study was supported by WHO, Geneva & ICMR, New Delhi LRO 8.6 Refinement of search filters to reduce information overload when retrieving evidence on variations in practice for NICE Quality Standards Craven J1 , Ayiku L1 , Barrett E1 , Walton L1 1 National Institute for Health and Care Excellence, United Kingdom Background: The National Institute for Health and Care Excellence (NICE) Quality Standards are concise sets of prioritised statements designed to drive measurable quality improvements within a particular area of health or care in the United Kingdom (UK). A central aim of Quality Standards is to reduce variations in practice in care that arise from non-adherence to guidance, and which result in poor quality care. In 2012, to support the development of Quality Standards, information specialists within the guidance Information Services (gIS) team at NICE developed a search filter to identify literature on variations in practice from bibliographic databases including MEDLINE, Embase, PsycINFO, and the Health Management Information Consortium (HMIC). Having used the filter over a period of time, systematic reviewers from the Quality Standards team reported that a large number of irrelevant references were being retrieved by the filter. In response, the information specialists undertook a project to increase the precision of the filter. The revised filter was then piloted for a Cochrane Database Syst Rev Suppl 1–327 (2015) 117 period of eight months until March 2015. Objectives: To improve the precision of the ‘variations in practice’ search filter without excluding key papers. Methods: 1. Analysis undertaken of the irrelevant results from previous searches to identify which aspects of the filter could be modified. 2. Modification of the filter, and assessment of its performance in retrieving key papers that had informed previous Quality Standards topics. 3. Piloting the revised filter on new Quality Standards topics to assess its ability to retrieve key papers and reduce the volume of irrelevant references. Results: Precision has been improved: the number of irrelevant references retrieved has been reduced to an acceptable level. Conclusions: Information overload through retrieval of irrelevant references has been reduced, saving the sifting time of systematic reviewers. The modified filter has proved to be robust in identifying key papers and will be used for all forthcoming Quality Standards variations in practice searches. Oral session 10 Filtering the information overload for better decisions O 10.1 The Embase project 2: crowdsourcing citation screening Noel-Storr A1 , Dooley G2 , Glanville J3 , Foxlee R4 1 Cochrane Dementia and Cognitive Improvement Group, Oxford University, United Kingdom; 2 Metaxis Ltd, United Kingdom; 3 York Health Economics Consortium, United Kingdom; 4 Cochrane Editorial Unit, United Kingdom Background: The Embase project has been managed since April 2013 by a consortium made up of Metaxis Ltd, the Cochrane Dementia and Cognitive Improvement Group and the York Health Economics Consortium. It uses a novel crowdsourcing method to screen citations. Objectives: To evaluate the effectiveness of using crowdsourcing to identify unique reports of randomised trials in Embase and to submit those reports to Cochrane’s Central Register of Controlled Trials (CENTRAL). Methods: We recruited a crowd to screen the search results identified from the monthly sensitive searches run in Embase (via Ovid SP). Using a bespoke online citation screening tool the crowd classify citations as ‘RCT/CCT’, ‘Reject’, or ‘Unsure’. Main outcome measures are performance of the crowd in terms of collective classification accuracy, quantity screened, timeliness, and screener recruitment, retention and engagement. Results: To date (March 2015) over 950 people have signed-up to take part in the project and over 120,000 citations have been collectively screened by the crowd. The results for four independent validation studies (two completed; two ongoing) to assess crowd accuracy will be presented. The two completed validations show crowd sensitivity of 99.8 and 99.9, and specificity of 99.8 and 99.7. The reference standard used in both cases was determined by expert screeners. Conclusions: This new approach to screening has brought significant efficiencies to trial identification. Crowdsourcing has proved to be both feasible in terms of recruitment and robust methodologically. It has meant that more trials have been identified more quickly, making CENTRAL a more valued repository of trial reports. O 10.2 A machine learning approach to classifying articles for literature reviews based on titles and abstracts Steeves S1 , Buchanan-Hughes A1 , Leonard S1 1 Costello Medical Consulting Ltd, United Kingdom Background: Machine learning could improve the efficiency of conducting literature reviews by automating review of titles and abstracts for relevant articles. Objectives: This study investigated if it is feasible to use the Fisher classification method to distinguish between relevant and irrelevant articles for literature reviews based on their titles and abstracts, and to determine if success varies depending on the type of classification being performed. Methods: Datasets were created from abstract lists from three systematic reviews. To explore if the algorithm performed better on particular types of classification (by study design, by outcomes measured, by interventions used or by disease area), decisions at various points of the eligibility flowcharts Table 1 Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Cochrane Database Syst Rev Suppl 1–327 (2015) 118 were tested separately. Articles were labelled as ‘relevant’ or ‘irrelevant’ at each of these stages. The datasets were processed to remove duplicates and to adjust for imbalances in ‘relevant’:‘irrelevant’ abstracts as a possible confounder. Articles with only a title or only an abstract were retained. Each dataset was divided into training (60%), cross-validation (20%) and test sets (20%). Accuracy was measured using classification accuracy and the F2 score which favours correct classification of ‘relevant’ items. After training the classifier algorithm, we optimised the F2 score on the cross-validation set by adjusting the thresholds at which a ‘relevant’ or ‘irrelevant’ label was assigned. Items falling below these thresholds were marked as ‘uncertain’ by the algorithm and excluded from the F2 score calculation. The final F2 score was calculated on the test set. Results: Classifiers were trained on six datasets with final F2 scores varying from 0.663 to 0.879 (Table 1). Classifications by study type and disease area outperformed those based on outcome measures or interventions. Conclusions: The Fisher classification method was successful at classifying relevant and irrelevant articles based on their titles and abstracts, particularly for classifications based on study design and disease area. We intend to investigate the performance of other machine learning algorithms on this task. to complete, with fewer resources, are needed. Objectives: To evaluate the performance of a new natural language processing (NLP) algorithm. Methods: We developed a new NLP algorithm based on diverse relevance ranking models for MEDLINE citations. A linear combination of two ranking scores from semantic relevance ranking and latent Dirichlet allocations was used to predict an overall relevance ranking score for each citation. To evaluate the performance of this new method, we selected a convenience sample of five SRs published by Cochrane. We estimated area under curve (AUC), sensitivity, false-positive rate, total screening burden, and percentage of reduction in screening. We compared the new pooled effect size to the published one using the Altman and Bland method. Results: The new NLP algorithm achieved an average AUC of 0.82 (range: 0.49 to 0.95). With 70% reduction of the number of citations to be screened, we observed over 80% sensitivity in four out of five SRs. We did not find significant difference between the published effect size and the new pooled effect size even after 90% reduction of citations. Conclusions: NLP algorithms showed promising results on accelerating the SR process and reducing workloads. Future work is needed to expand the search beyond MEDLINE and validate this pilot study. O 10.4 O 10.3 Natural language processing to increase efficiency of systematic reviews: a pilot study Murad MH1 , Wang Z1 , Li D1 , Liu H1 1 Mayo Clinic, USA Background: Systematic reviews (SRs) require significant resources and time to complete. A typical SR takes between six and 18 months. Interventions that can make SRs swifter Evaluation of the priority ranking capabilities of SWIFT (Sciome Workbench for Interactive, Computer-Facilitated Text-mining) software Walker V1 , Holmgren S1 , Thayer K1 , Rooney A1 , Macleod M2 , Currie G2 , Sena E2 , Sherratt N2 , Rice A3 , Howard B4 , Shah R4 , Pelch K1 1 National Institute of Environmental Health Sciences (NIEHS)/National Institutes of Health (NIH), USA; 2 Centre for Clinical Brain Sciences, University of Edinburgh, Scotland; 3 Department of Surgery and Cancer, Imperial College, England; 4 SciOme LLC, USA Table 1 Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Cochrane Database Syst Rev Suppl 1–327 (2015) 119 Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Cochrane Database Syst Rev Suppl 1–327 (2015) 120 Background: There is growing interest in assessing the ability of machine learning approaches to priority rank studies as a way to reduce the human burden in screening literature when conducting a systematic review. Objectives: To assess the performance of Sciome Workbench for Interactive, Computer-Facilitated Text-mining (SWIFT) priority ranking algorithm to identify studies considered relevant based on manual screening. Methods: Four case studies representing a range of complexity and size were used to assess the performance of SWIFT: (1) transgenerational inheritance of disease, (2) bisphenol A (BPA) and obesity (3) perfluorooctane sulfonate/perfluorooctanoic acid (PFOS/PFOA) and immunotoxicity, and (4) neuropathic pain. For each case study two independent reviewers manually screened results to determine study relevance and identify test sets of 30 to 400 included and excluded references. The test sets were used to priority rank the literature search results in SWIFT for relevance using an algorithm that considers term frequency (title, abstract, MeSH headings and SuppChem annotations) and Latent Dirichlet Allocation (LDA) topic modeling. This ranking was evaluated with respect to (1) number of studies that needed to be screened in order to identify 90% and 95% of known relevant based on manual screening, and (2) the ’Work Saved over Sampling’ (WSS) performance metric, which defines, for a specific level of recall, the percentage reduction in effort achieved by a ranking method compared to a random ordering of the documents. Results: For all four datasets, using 100 training examples and LDA topic modeling, the prioritization procedure reduces the number of citations that must be screened to achieve a recall rate of 90% (Table 1) by 50% or more. For the more stringent recall rate of 95%, the range in the number of citations screened was reduced by 44%, for the neuropathic pain dataset, to 80%, for the PFOS/PFOA dataset. The greatest increases in screening efficiency were observed in the more targeted topics. Conclusions: Text-mining and machine learning programs such as SWIFT can be valuable tools to reduce the human screening burden. Oral session 11 Rapid reviews and overviews of reviews O 11.1 An evidence map of studies evaluating methods for conducting, interpreting, and reporting overviews of systematic reviews Lunny C1 , McKenzie J1 , McDonald S1 , Lunny C1 , Brennan S1 , Lunny C2 1 Australasian Cochrane Group, Australia; 2 Australasian Cochrane Group, Monash University, Australia Background: Methods for conducting, interpreting, and reporting overviews of systematic reviews are in their infancy. Where possible, guidance for overviews should be based on Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. methods that have been evaluated and shown to have better performance. It is timely to map the methods literature on overviews, so as to determine what methods research should be undertaken as a priority. Objectives: 1. To populate a framework with methods that have, or may be used, in conducting, interpreting, and reporting overviews of systematic reviews. 2. To create an evidence map of studies that have evaluated these methods. Methods: The research will be undertaken in two stages to meet objectives 1. and 2. We plan to search methods-specific databases (Cochrane Methodology Register, Meth4ReSyn library, AHRQ Effective Health Care Program), in addition to a general search for overviews in MEDLINE. Stage I: Methods that have been used in overviews will be identified from descriptive cross-sectional studies on methods used in overviews; guidelines; and editorials. We will extract data on: the type of paper (descriptive cross-sectional paper; guideline; editorial); description of the method; noted advantages/disadvantages; and where the method is located in the review process. Stage II: Evidence of the performance of methods will be identified from systematic reviews of methods studies, and methods studies. Evaluations will be described narratively, and mapped to the framework of methods identified in stage I. Results: The results will be presented at the Colloquium. Conclusions: The results of this mapping process will be useful for cataloguing overview of systematic review methods, informing guidance, and identifying and prioritising methods research in this field. O 11.2 Rapid review programs to support healthcare and policy decision making: a descriptive analysis of processes and methods Polisena J1 , Garritty C2 , King V3 , Stevens A2 1 Canadian Agency for Drugs and Technologies in Health, Canada; 2 Ottawa Hospital Research Institute, Canada; 3 Oregon Health Sciences University, USA Background: Healthcare decision makers often make decisions under limited timeframes that preclude the completion of more comprehensive evidence synthesis. Rapid reviews (RRs), using streamlined systematic review methods, are used frequently for evidence synthesis to support such decisions. Objectives: Our primary objective was to describe the processes and methods used across international programs to produce RRs. We also sought to understand the underlying themes associated with these programs and to identify research opportunities in rapid review methods and impact. Methods: We contacted representatives of healthcare RR programs to inquire about their methods and processes. The characteristics of each program’s processes and methods were summarized and compared to highlight potential themes and trends related to the practice of RRs. In addition, knowledge gaps to inform research priorities in RRs were identified. Results: Twenty-nine RR programs representing academia, government, research institutions, Cochrane Database Syst Rev Suppl 1–327 (2015) 121 and not-for-profit organizations participated in our survey. Responses revealed that the main objectives for RRs were to inform decision making on funding healthcare technologies, services and policy, and program development. Central themes influencing the process and methods used were timelines, available resources, the complexity and sensitivity of the research topics, and proprietary nature of the product. Conclusions: Observed differences in processes and methods across programs may result from the continuous development of RR methods, customization of RRs for decision makers, and definition of ‘rapid’ by organizations. The primary research priority is to develop a typology of RRs to understand the strengths and limitations of various RR forms better and the level of synthesis performed. Future research also should investigate the impact of RR methods and reporting to support informed healthcare decisions, the effects of potential biases that may be introduced with streamlined methods, and the effectiveness of RR reporting guidelines on transparency. We selected 32 pairs of documents. The preliminary analysis showed no serious mismatching. The final results will be presented at the Colloquium. Conclusions: The timeframe to produce evidence is becoming shorter. It is critical to determine if ultra-rapid evidence summaries produced by HTA trained teams are reliable. O 11.3 Background: Treatments that are matched to patient risk (stratified care) have the potential to improve effectiveness of primary care for patients with musculoskeletal pain. However musculoskeletal pain conditions are extensive and the knowledge base is large. To inform development of matched treatments a rapid yet detailed evidence summary on effectiveness of available treatment options was needed. Objectives: To develop an approach to synthesizing large evidence summaries; rapidly to synthesise and appraise current best evidence on treatment options for five most common musculoskeletal pain presentations in primary care and to summarise the available evidence on treatments for patient risk subgroups using stakeholder groups. Methods: Evidence synthesis followed a pyramidal approach using national clinical guidelines, policy documents, clinical evidence pathways and summaries as starting point. Recommendations on available treatment options for shoulder, neck, knee, back and multisite pain were extracted consecutively. Systematic searches of bibliographic databases were conducted to identify and retrieve additional published trials that had not yet been summarised or where evidence gaps existed. Quality of evidence was assessed based on modified GRADE quality ratings and strength of evidence. Evidence summaries were subsequently presented to stakeholders (including health service managers, clinicians and researchers) for interpretation and to identify appropriate treatment options that might be matched to patient risk subgroups. Results: Via a rapid yet systematic and comprehensive approach, pragmatic summaries of the evidence base on treatment options for five musculoskeletal pain presentations were completed. Based on current best evidence, identification of matched treatment options according to patient risk subgroups appears feasible across musculoskeletal pain presentations. Conclusions: Effective healthcare delivery and clinical practice depend on high quality evidence which provides depth as well as breadth of coverage. A novel and pragmatic approach to rapid synthesis of large evidence is explored and found applicable within musculoskeletal pain field. Are 3 days enough to capture the key evidence for HTA documents? Ciapponi A1 , Bardach A1 , Glujovsky D1 , Rey Ares L1 , Garcı́ a Martı́ S1 1 Instituto de Efectividad Clı́ nica y Sanitaria - IECS, Argentina Background: Rapid reviews have emerged as an efficient approach to synthesizing evidence, within a four to five week timeframe, for informing decision makers in health care settings. It is uncertain whether even very shorter timeframes that are needed sometimes in specific healthcare decisions, e.g. two to three days elaboration-time, would still be adequate for capturing the key evidence that forms part of more elaborate Health Technology Assessment (HTA) documents. Objectives: To compare the conclusions and analyze the amount and direction of the evidence included in HTA documents produced in an ultra-rapid way compared to the more normal four to eight weeks. Methods: IECS is an Argentinean HTA agency that produces both of the aforementioned types of documents according to the urgency and needs of decision-makers. The documents are based on focused search strategies in meta-search engines and online biomedical literature databases, to identify systematic reviews, clinical practice guidelines, HTAs, coverage policies, and selected primary research. The ultra-rapid HTAs are prepared by highly-trained staff who select the most important evidence according to their own judgement. The ’slower’ HTAs allow a more exhaustive assessment of the evidence. We selected pair of documents, one done in two days and the other in five weeks, oriented to the same research question. The longer document needed to be published within a year of the shorter one. The additional evidence identified by the newer document, which was compiled at a later date than the ultra-rapid HTA, was excluded and the conclusions modified wherever necessary. Pairs of independent researchers extracted the outcomes, and disagreements were solved by a third researcher. Results: Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. O 11.4 Effective treatment options for musculoskeletal pain conditions: a rapid meta-synthesis of current best evidence within primary care Babatunde OO1 , Jordan JL1 , Van der Windt DA1 , Foster NE1 , Protheroe J1 , Working Group S1 1 Research Institute for Primary Care and Health Sciences, Keele University, Keele, Staffordshire, ST5 5BG, United Kingdom Cochrane Database Syst Rev Suppl 1–327 (2015) 122 Oral session 12 Prediction models O 12.1 CHecklist for critical Appraisal and data extraction in systematic Reviews of clinical prediction Modelling Studies (CHARMS) Moons K1 , de Groot J1 , Bouwmeester W1 , Vergouwe Y1 , Mallett S2 , Altman D2 , Reitsma J1 , Collins G2 1 Julius Center for Health Sciences and Primary Care, The Netherlands; 2 University of Oxford, United Kingdom Background: Publications on multivariable clinical prediction models have become abundant for both prognostic and diagnostic purposes. Systematic reviews of these studies are increasingly required to identify and critically appraise the existing evidence. There is currently no checklist or tool providing guidance for systematic reviews of studies developing or validating prediction models that can assist reviewers to define the review objectives and appraise study methodology. Objectives: To develop a checklist to help reviewers framing a well-defined review question, and to determine which details to extract and critically appraise from primary studies on the development or validation of multivariable diagnostic or prognostic prediction models, with a view to assessing the risk of bias and sources of heterogeneity. Methods: We critically examined existing reporting guidelines and quality assessment tools, key methodological publications on clinical prediction modelling, and tools used in published systematic reviews of multivariable prediction models, to identify the relevant characteristics and domains. The checklist was tested in various systematic reviews. Results: We identified seven items important for framing the review question (diagnostic versus prognostic model, intended scope of the review, type of prediction modelling studies, target population, outcome to be predicted, time span of prediction, intended moment of using the model), and 11 domains to critically appraise the primary included studies (source of data, participants, outcome, predictors, sample size, missing data, model development, model performance, model evaluation, results, interpretation). Both were combined into the CHecklist for critical Appraisal and data extraction for systematic Reviews of prediction Modelling Studies (CHARMS). Conclusions: CHARMS is designed to assist reviewers to help systematic Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. reviewers when framing their review objectives, and to determine which data to extract and critically appraise from primary studies on the development and/or validation of (diagnostic and prognostic) prediction models. O 12.2 Individual Participant Data (IPD) meta-analysis of diagnostic and prognostic modeling studies: a practical introduction to their rationale and conduct Debray T1 , Riley R2 , Rovers M3 , Reitsma J1 , Moons K1 1 Julius Center for Health Sciences and Primary Care, The Netherlands; 2 Keele University, United Kingdom; 3 Radbound University Nijmegen Medical Center, The Netherlands Background: The development and (external) validation of diagnostic and prognostic prediction models is an important area in contemporary medical research. During the past few years, evidence synthesis and meta-analysis of individual participant data (IPD) has become increasingly popular, not only for intervention research but also for improving the development, validation and generalizability of diagnostic and prognostic prediction models. IPD metaanalyses (IPD-MA) provide unique opportunities to improve development and enhance the applicability of prediction models across (sub)populations and settings. There is, however, little guidance on how to conduct an IPD-MA to develop and validate diagnostic and prognostic prediction models, or how to interpret their findings. Objectives: To describe how IPD-MA of diagnostic and prognostic modeling studies differ from IPD-MA for assessing treatment effects. Methods: We identify key advantages and challenges in IPD-MA of prediction models. Subsequently, we provide recommendations for the design of such IPD-MA including the selection of relevant studies, Finally, we discuss statistical methods for handling between-study heterogeneity, missing data, and other issues regarding prediction model development and validation. We illustrate all these concepts using various empirical examples across medical disciplines. Conclusions: The guidance provided in this work may help meta-analysts in prediction modeling research to decide upon appropriate strategies when conducting an IPD-MA, and assist readers, reviewers and practitioners when evaluating the quality of resulting evidence. Cochrane Database Syst Rev Suppl 1–327 (2015) 123 Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Cochrane Database Syst Rev Suppl 1–327 (2015) 124 O 12.3 Imputation of missing adjusted data for non-significant factors in meta-analysis of prognostic studies Wang L1 , Kennedy S2 , Romerosa B3 , Kwon H4 , Kaushal A2 , Craigie S1 , Chang Y2 , Almeida C5 , Izhar Z2 , Couban R1 , Parascandalo S6 , Guyatt G2 , Reid S2 , Khan J2 , McGillion M2 , Busse J1 1 Michael G DeGroote Institute for Pain Research and Care, McMaster University, Canada; 2 McMaster University, Canada; 3 Complejo Hospitalario de Toledo, Spain; 4 Wayne State University, USA; 5 Federal University of Rio Grande do Sul -UFRGS, Brazil; 6 University College Cork, Republic of Ireland Background: Many prognostic studies use data-driven models: each independent factor is tested in a bi-variable analysis, and only those that show evidence of association (e.g. P value ≤ 0.05) are entered into an adjusted model. Others report data only for those independent factors in their adjusted model that show a significant association with the dependent variable. Systematic exclusion of these data presents a risk of overestimation by only Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. pooling estimates of association from predictors that appear in adjusted regression models, and for whom data is provided. Objectives: To investigate if imputation of missing non-significant data in final regression model will avoid overestimation of the predictive power of the risk factors, using predictors for persistent pain after breast cancer surgery as an example. Methods: We pooled nine predictors to explore their association with the development of persistent pain following breast cancer surgery using random-effects models. We imputed an odds ratio (OR) of ‘1’ for predictors that were excluded from adjusted analyses due to non-significant bi-variable analyses, or that were reported, but with no data due to lack of significance in the final regression model. We acquired the associated variance for all such imputations using the hot deck approach. We performed sensitivity analysis to examine the impact of imputing data for non-significant predictors excluded from adjusted analyses by re-running our analyses and excluding the imputed data. Results: Fifty-nine (51.3%) out of 115 study-sets for nine predictors failed to reported the adjusted data for non-significant predictors. Our sensitivity analyses found no significant differences in results whether or not we incorporated missing data for non-significant predictors Cochrane Database Syst Rev Suppl 1–327 (2015) 125 (Table 1). However, the associations of the nice predictors with persistent pain were consistently larger in meta-analyses based on the adjusted data only than in the full analyses including imputed missing data. Conclusions: Imputation of missing data for non-significant predictors did not cause any significant associations to lose significance, but the magnitude of association was reduced. O 12.4 Are predictions from test accuracy meta-analyses valid in practice? invalid a tailored estimate may be appropriate (Table 1). Furthermore the differences in estimates between the models could potentially affect patient management. Conclusion: Statistical validation of a test’s PPVs and NPVs could be part of the synthesis process of a test evaluation meta-analysis. Both the likely validity of the standard and tailored estimate in a new population could be ascertained at this stage. In the examples analysed, standard meta-analysis seldom yielded a valid estimate for the test’s performance. Importantly, this was often remedied by taking a tailored approach to the meta-analysis, which was more likely to yield a valid estimate. Willis BH1 , Riley RD2 University of Birmingham, United Kingdom; 2 University of Keele, United Kingdom Rapid oral session 19 Global health: Helping with evidence Background: Although meta-analysis may synthesize estimates for a test’s accuracy, heterogeneity often blights these and it is not always clear whether clinicians should trust these estimates when applying them to their own practice. Objectives: Apply a novel statistical method to determine whether the summary estimates of a test’s positive and negative predictive values (PPV and NPV) are likely to be valid in practice. Methods: Using four test accuracy reviews as examples, a univariate meta-analysis model was used to derive estimates for the likelihood ratios. Based on previously reported methods this model was extended to produce a tailored estimate specific to the setting of interest. The PPV and NPV for the tests were the main outcomes of interest. A new validation statistic, Vn is introduced that assesses the validity of the standard and tailored estimate for the PPV and NPV in a new population. This is derived using a cross-validation procedure that compares the observed and expected post-test predictions in an omitted study on multiple occasions. A significant Vn (P value < 0.05) suggests the estimate would not be valid in a new population. Results: For the four examples, each meta-analysis model derived four PPV estimates and four NPV estimates. Using the standard meta-analysis model, Vn was significant in seven of the eight estimates. In contrast, for tailored meta-analysis, Vn was significant in only three of the eight estimates suggesting that when a standard estimate is likely to be RO 19.1 1 Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. From Dublin to Vienna - the Cochrane Waltz Larun L1 , Gundro Brurberg K1 , Odgaard-Jensen J1 1 Norwegian Knowledge Centre for the Health Services, Norway Background: The time from protocol to review was around two years (median) in 2014. Cochrane Review authors have access to online resources, workshops and courses in addition to the Cochrane Handbook and the group’s specific resources and a Cochrane Author Support Tool to be published. This should enable the review authors to waltz through the review process (Waltz definition 1: to succeed easily or breeze through). Why is it that some authors feel they are stuck in a true Viennese waltz (Waltz definition 2: a complex dance that consists only of multiple turns and changes in step)? Objectives: How to keep review authors in the loop? Methods: Case study including three professional systematic review authors experienced in writing Cochrane Reviews. Data collected from Archie, emails and personal memory analysed narratively. Results: Review authors are a vital part in the production of Cochrane Reviews and we should all take care that they feel valued. Review authors are asked to submit revised versions within three months and this should apply to editorial processes as well. We should consider Cochrane Database Syst Rev Suppl 1–327 (2015) 126 specific editorial support for non-native English speakers to facilitate clear writing. Review authors should be made aware of likely timelines in both absolute months and elapsed time from title to protocol and protocol to review to set clear expectations. Conclusions: Review authors, like users, consumers or customers might have valuable insight into pitfalls and enhancers for how to keep reviewers and how to get people on board. Review authors should always be informed about the viability of producing a Cochrane Review in relation to existing workload so they can make informed decisions. RO 19.2 Cochrane protocols: strategies for improving production Marcus S1 , Noel-Storr A1 1 Dementia and Cognitive Improvement Group, United Kingdom Background: Protocols describe the rationale and methodological framework for a subsequent systematic review and must be published before the review. Whilst they usefully describe the need to answer a valid research question, they are time consuming for authors and can take several months or even years to publish. The peer referee and editorial process delay publication further. Objectives: 1. To assess protocol generation times within the Cochrane Dementia Group. 2. Trial new method/s in protocol production within the context of a large NIHR funded programme grant. 3. Propose other possible models of protocol production. Methods: 1. We performed a cross-sectional analysis based on data in Archie of our ten most recent intervention protocols to see how long they took from registration to publication. 2. For a suite of 12 reviews on Modifiable Risk Factors we used a generic protocol written in-house with more input than usual from the editorial base –(i) consultant methodologist; (ii) designated systematic reviewer (iii) greater use of our group’s specialist/contact editors; (iv) one peer referee. Results: 1. Times taken from registration to publication varied from 32 to 327 weeks. Mean time: 109 weeks. 2. We will report on time taken from registration to first draft, and to publication. We will report what worked well and what remained a challenge. Conclusions: Both methodologies are time consuming, but option 2 has potential for reducing registration to publication time. However a third option might be to publish the protocol as an appendix to the full review. This could incorporate method 2 above, obviating the need for sequential publication. It could considerably reduce time from registration to review publication by reducing the protocol editorial process. Additionally a generic protocol could increase consistency across reviews. Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. RO 19.3 A new approach to ‘empty’ reviews Kew KM1 , Welsh EJ1 1 Airways, United Kingdom Background: Cochrane review groups are encouraged to prioritise reviews that are likely to include studies over those that might not, but questions that identify no trials can be important to people making healthcare decisions. ‘Empty’ reviews based on comprehensive literature searching, especially those that have been identified in prioritisation exercises, can confirm and highlight important research gaps and inform the design of future trials. Within Cochrane ‘empty’ reviews are subject to the same authoring and editorial requirements as reviews with included studies. This conflicts with the founding principle of Cochrane and the strategy to 2020 which encourages prudent use of resources. Objectives: To develop and evaluate new approaches to ‘empty’ reviews Methods: We will collect data about the time taken for protocol and review development of recent ‘empty’ Airways Group reviews. We will also look at our review proposals’ log to assess the number of times titles were deemed clinically important but rejected because no trials were anticipated. Results: We will put forward alternative approaches to ‘empty’ reviews. Our criteria will be: 1. maintains methodological integrity up to the point that a review can be confirmed empty; 2. minimises the workload for authors, editorial staff, peer reviewers and copy-editors; 3. publicises an important research gap quickly and succinctly to funders and researchers. The approaches we describe aim to prevent research waste and are not intended for non-priority questions. Conclusions: ’Empty’ reviews for important research questions can be a useful tool to confirm and publicise research gaps. RO 19.4 Stop press! Experience of managing serious complaints about a Cochrane Review Armstrong S1 , Jordan V2 , Farquhar C2 1 Cochrane Menstrual Disorders and Subfertility Group, United Kingdom; 2 Cochrane Menstrual Disorders and Subfertility Group, University of Auckland, New Zealand Background: Serious complaints about Cochrane Reviews are rare, but can pose a challenge to authors as to how best to respond. Cochrane has a webpage dedicated to submitting a complaint, however it doesn’t publish guidance on how to address complaints. Should authors seek advice beyond the scope of the review? Should complaints and responses be published? How much credence should authors give critics who have an obvious but undeclared conflict of interest? Objectives: To offer a personal perspective on managing serious criticism of a Cochrane Review. Methods: An account of how a group managed a series of complaints about a Cochrane Review. Results: The initial complaint Cochrane Database Syst Rev Suppl 1–327 (2015) 127 letter contained diverse criticisms including the title, the primary outcome and the use of the term ‘cell-tracking algorithms’ amongst others. There was agreement that we had robust evidence to defend our protocol and to not make any changes. We sought advice from an expert researcher and clinician in embryology, who supported our conclusions. In the meantime, the full review had been published, and we decided to republish the review with both letters appended. A second letter was received one month later. It condemned the use of the intention-to-treat principle, the use of miscarriage per randomized woman and the heterogeneity of definitions of miscarriage between studies. This letter concluded with the demand for the review to be withdrawn and republished under different authorship. Again, we sought expert advice and referenced statistical advice from the Cochrane Menstrual Disorders & Subfertility Group and CONSORT. A reply was sent that defended our methodological decisions. An internet search of the critics revealed that they both have undisclosed links with the time-lapse imaging industry. Conclusions: Seeking advice from methodological, statistical and content experts is useful to help clarify the complaints raised and determine how to respond. Commercial conflicts of interest may not be disclosed by critics, but it is important to respond equitably and to publish all correspondence with the review. RO 19.5 Discernment of academic conflicts of interest for Cochrane Systematic Reviews Miyazaki C1 , Ota E1 , Mori R1 , Sasaki H1 , Kirkham J2 , Dwan K3 1 Department of Health Policy, National Center for Child Health and Development, Japan; 2 The University of Liverpool, UK; 3 Cochrane Editorial Unit, UK Background: Cochrane has set a ground policy to ensure that its systematic reviews are free of conflict of interest (COI) and to prevent potential influence from commercial sponsors or sources. Although Cochrane’s declaration of interest focuses largely on avoiding potential financial COI, guidance on academic COI remains ambiguous. Health policy promotion occurs through gathering evidence of healthcare interventions from systematic reviews, and involves clinical experts and consensus from academic conferences. The extent to which the evidence is influenced by academic and political interest, and the career driven incentive of the authors, is not known. Apart from disclosures of a financial nature, academic interest may be the driving force for drawing a conclusion. There are many research groups and institutes that have different standards in their code of conduct policy, therefore, there is an implication of a lack of consistency in practising the elimination of non-financial COI. Objectives: To conduct a survey of the existing guidance on COI management within the 54 Cochrane Review Groups and to characterize academic COI. Methods: An internet-based survey will be used to acquire any statement of academic COI policy from the 54 Cochrane Review Groups. The survey will identify whether a comprehensive definition of academic Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. COI exists. Review groups without a disclosed academic COI will be selected for assessment, by relating review authors to authors of included clinical trials in published Cochrane Systematic Reviews. Results: The results will show how many Cochrane Systematic Review authors have conducted their own clinical studies and the Cochrane Review Groups in which affiliations are common. Conclusions: This research will present the underlying concern about the need to include independent people in review teams to exercise avoidance of academic COI, retain impartiality in the outcome judgments of the reviews and to promote well-balanced review teams. RO 19.6 Selective reporting of outcome data: are we following the Cochrane Handbook guidance? Rutjes AW1 , Maione A1 , Silletta MG1 , Di Nisio M1 1 Centre for Systematic Reviews, Fondazione Universita’ ‘‘Gabriele d’Annunzio”, University of Chieti-Pescara, Italy Background: The Cochrane Handbook advises to judge the risk of bias (RoB) potentially introduced by selective reporting of outcome data (SORB) on the within-trial level, while judging if the outcomes expected to be reported in a given field are addressed. Objectives: To summarize and discuss applied methods to judge and handle RoB due to selective reporting in Cochrane Reviews. Methods: We obtained unique reviews from the Cochrane Musculoskeletal Group (CMSG) and the Cochrane Peripheral Vascular Diseases Group (PVD) published in 2011 to 2015, including at least 1 RCT. Extraction items included definitions used to judge the RoB by selective reporting, the frequency of scoring low, high or unclear RoB due to selective reporting, and the handling of this item in the ‘Summary of findings’ tables (SoF). Results: Complete results will be presented at the Colloquium. Up to 2013, 162 reviews were identified, of which 42 PVD and 24 CMSG reviews were considered. Figure 1 shows that 83% of the reviews considered SORB. Of these, the majority of reviews referred to the Cochrane Handbook for the operationalization of how to assess SORB Figure 1 Cochrane Database Syst Rev Suppl 1–327 (2015) 128 Table 1 (Table 1). Most reviews assess SORB at the within-trial level, matching outcomes reported in protocols, trial registrations and methods sections to outcomes addressed in the results section. Only a minority (9%) followed the Handbook guidance to assess SORB at the within-trial and topic level; 17% focused on completeness of reporting of review-defined outcomes (review outcome level). SoF tables were provided in 44% of the included reviews (Figure). SORB was infrequently (15%) mentioned as a reason to downgrade the confidence in the estimate of the effect. Conclusions: Cochrane guidance on SORB is insufficiently followed by CMSG and PVD reviewers. SORB is typically assessed on the within-trial level, without considering whether at least the main outcomes relevant to the topic or review question are addressed across trials. SORB seems to be given insufficient attention while GRADE-ing our confidence in the estimates of the effects. We suggest a stakeholder meeting to evaluate and discuss hurdles to assess SORB and to improve methods. RO 19.7 Developing core competencies for scientific editors of biomedical journals Galipeau J1 , Moher D2 , Shamseer L2 , Barbour V3 , Bell-Syer S4 , Cumpston M5 , Deeks J6 , Garner P7 , MacLehose H8 , Straus S9 , Tugwell P10 , Wager E11 , Winker M12 1 Ottawa Hospital Research Institute, Canada; 2 Ottawa Hospital Research Institute; University of Ottawa, Canada; 3 Committee On Publication Ethics (COPE), Australia; 4 Cochrane Wounds Group, UK; 5 Cochrane Central Executive, Australia; 6 Cochrane DTA Working Group, UK; 7 Cochrane Infectious Diseases Group, UK; 8 Cochrane Editorial Unit, UK; 9 University of Toronto, Canada; 10 Cochrane Musculoskeletal Group, Canada; 11 Sideview, UK; 12 World Association of Medical Editors (WAME), USA Background: Scientific editors play a key role in the assessment, acceptance, and dissemination of high quality reports of medical research. However, as a group they are generally not well educated in their position. This may be, in part, due to a lack of evidence of what makes a good scientific editor or how to train them properly. The result is that there are no standard criteria for determining one’s suitability for the position or training needs; in short, we don’t Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. know what competencies are essential for effective scientific editing. Objective: The objective of this project is to develop a globally accepted minimum set of core competencies for scientific editors (i.e. those responsible and accountable for selecting the scientific content of a peer-reviewed biomedical journal), including Cochrane editors. Methods: Phase 1 involves searching published and unpublished literature for information relating to competencies of scientific editors; the output will be a scoping review. Phase 2 consists of a training needs assessment to ascertain the needs of scientific editors from diverse regions and varying journal types and sizes. Phase 3 is a Delphi process open to all scientific editors globally; the output will be a list of competencies deemed important for performing the role of scientific editor. Phase 4 features a consensus meeting to achieve agreement on a minimum set of competencies for scientific editors. Phase 5 involves the development of a globally accessible curriculum, training modules, and certification program. Discussion: We will provide an update on our progress and will share available results from the first three project phases. The goal of the project is to provide enhanced credibility and consistency to the role of scientific editor across the spectrum of biomedical journals. Funded in part by the Cochrane Central Executive as a high priority issue, the project is international in scope and includes several key stakeholder groups, such as the World Association of Medical Editors (WAME), the Committee on Publication Ethics (COPE), and Cochrane. RO 19.8 Updating patient information: a literature review Wegmann M1 , Will R1 , Ehrlich M1 , Janßen I1 1 IQWiG, Germany Background: The IQWiG produces evidence-based health information (HI) that is updated every three years. Maintaining a fixed update cycle requires a major investment of resources and does not always appear to provide the most suitable timeframes for some of the information. Objectives: (1) Which solutions does the current literature offer for updating HI? (2) How do other international organisations deal with this issue? (3) Are there recommendations or suggestions that can be meaningfully applied for our purposes? Methods: A systematic search of the literature accompanied by supplementary handsearch was performed. Any Cochrane Database Syst Rev Suppl 1–327 (2015) 129 publications offering data relevant to our objectives were included. The search strategy design was created by two reviewers and the subsequent search was carried out by the same two reviewers. Two people independently viewed the titles and abstracts. Relevant titles were ordered as full articles. One reviewer extracted the data from these articles with a second checking the work. Organisations also involved in the production of HI were contacted via email. Results: The search produced 2324 hits. Following title and abstract screening 52 references remained. Following full-text assessments, 25 references were included. Six of the organisations we contacted responded to our request. Conclusions: (1) The current literature does not provide any suggestions concerning optimal update cycles or strategies for HI. (2) Interviews with the organisations showed that they have not implemented a standardised procedure for updating, nor do they have a fixed cycle for updates. Some of the organisations use ’horizon scanning’. (3) The literature identified confirms that regular updates are necessary and that marking information with a recent date sends an important message to the reader. The use of statistical methods to calculate the influence of new study results on meta-analysis effect estimators is highly time-consuming. Nevertheless, this would be a way of identifying HI with stable effects that would remain up-to-date longer. New techniques such as semi-automated title and abstract screening or abbreviated searches may be viable for our updating process. 249 recommendations and supporting evidence summaries for antithrombotic therapy, we performed user-testing with 12 physicians in Norwegian hospitals. Our multidisciplinary team of designers, programmers, clinicians and researchers applied qualitative methods with both direct observation and semi-structured interviews and performed iterative improvements in functionality of MAGICapp. Results: We implemented improvements in MAGICapp related to the following main issues in accessibility: delay from login to accessing content (e.g. more than three minutes on old browsers on hospital desktop stations); need for a comprehensive disclaimer for use in practice, suboptimal display of recommendations and PICO questions (e.g. physicians did not intuitively find where to click to access deeper layers of information); poor scrolling functionality on mobile devices; too much screen space used for less relevant information (e.g. table of contents); need for better offline access, and search functionality for recommendations. Problems with understanding and utility of content related to: conceptual understanding of GRADE (e.g. quality of evidence, risk and uncertainty), wording being sometimes imprecise or technical, incomplete access to references and abstracts. Conclusions: The barriers identified illustrate challenges and solutions for getting evidence into practice through innovative technological platforms. Rapid oral session 20 Disseminating evidence RO 19.9 Getting best current evidence into practice through a web-based authoring and publication platform: qualitative study of technological barriers and solutions Brandt L1 , Kristiansen A1 , Agoritsas T2 , Fog Heen A1 , Guyatt G3 , Vandvik PO4 1 Department of Medicine, Innlandet Hospital Trust-division Gjøvik, Norway; 2 Department of Clinical Epidemiology and Biostatistics, McMaster University, Canada, Canada; 3 Department of Clinical Epidemiology and Biostatistics, McMaster University, Canada; 4 Institute for Health and Society, Faculty of Medicine, University of Oslo, Norway Background: Trustworthy evidence summaries and guidelines are key information resources for clinicians. Digital structuring of content and optimization of presentation formats for a variety of digital platforms can facilitate successful dissemination and use at the point of care. Objectives: To test and improve the accessibility, understandability and utility in meeting clinicians’ information needs of evidence summaries and recommendations published in multilayered formats on the web, smartphones and tablets. Methods: We have created a platform (www.MAGICapp.org) to author, publish and update dynamically trustworthy evidence summaries and guidelines developed with the GRADE system. Following publication of the first guideline through MAGICapp with Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. RO 20.1 The development of an evidence based chronic diseases management resource for health practitioners to use at the point of care Khalil H1 , Chambers H1 , Munn Z2 , Porritt K2 1 Monash University, School of Rural Health, Australia; University of Adelaide, Australia 2 The Background: There is a large gap between evidence and practice within health care, particularly within the field of chronic disease. To reduce this gap, and improve the management of chronic disease, a collaborative partnership between two schools within a large university and two industry partners (a large regional rural hospital and a rural community health centre) was developed to research the area. Objectives: The aim of the collaboration was to promote the development of translation science and the implementation of evidence-based health care in chronic disease with a specific focus on developing evidence-based resources that are easily accessed by clinicians. Methods: A working group consisting of members of the collaborating organisations and an internationally renowned expert reference group was formed. The group acted as a steering committee and were tasked to develop a taxonomy for the resources. In addition, a peer review process of all resources was established. A corresponding reference group, consisting of researchers and clinicians who are clinical experts in various Cochrane Database Syst Rev Suppl 1–327 (2015) 130 fields, was involved in the review process. The resources developed by the group include evidence summaries and recommended practices made available on a web-based database and accessed via subscription by clinicians and researchers worldwide. Results: As of early 2015, there were 147 new evidence summaries and 30 recommended practices detailing the best available evidence on topics related to chronic disease management including asthma, diabetes, heart failure, dementia and others. Training sessions and a newsletter have also been developed for clinicians within the node to enable them to use the content effectively. Conclusions: The successful development of the collaborative partnership and its evolution into producing a valuable resource for the translation of evidence into practice in the areas of chronic disease management will enable clinicians to access updated, pre-appraised summaries at the point of care. RO 20.2 Disseminating good evidence for Cochrane Systematic Reviews: practice and experience from China Li X1 , Yang G1 , Liang N1 , Wieland S1 , Liu J1 1 Centre for Evidence-Based Chinese Medicine, Beijing University of Chinese Medicine, China Background: In 2007, the Centre for Evidence-Based Chinese Medicine from Beijing began training volunteers to identify and translate citations and abstracts of (non)randomized controlled trials (RCTs/CCTs) published in Chinese journals to the CAM Field for submission to CENTRAL. In 2014, we started exploring translation of Cochrane Review abstracts and plain language summaries (PLS) into simplified Chinese to contribute to Cochrane’s translation strategy. We introduce our work to share our experiences in translation, recruiting, training, and managing translators in the hope of generating discussion about the skills for these projects. Translation of Chinese RCTs/CCTs: In the past eight years, we have enrolled and trained over 80 volunteers from Beijing University of Chinese Medicine to translate Chinese RCTs/CCTs, and have submitted 11,405 translated records to CAM Field. There are currently 15 active volunteers, and we plan to submit another 1500 records in 2015. The RCTs/CCTs are mostly from electronic searching for specific topics, and many are from the included/excluded studies in systematic reviews conducted by our centre. The volunteers are mainly 3rd, 4th and 5th year undergraduates recruited annually by email and orally. We train volunteers in software (ProCite) use, translation principles, and methodology. Translation of Cochrane Review abstracts/PLS: We have trained 30 centre members and submitted abstracts/PLS for three Cochrane Reviews as an initial practice in coordination with Cochrane’s Translations Coordinator, while the Chinese Cochrane Centre decides overall strategy. A list of reviews are being translated using Smartling, and we are recruiting and training more volunteers. Conclusions: Since 2007 we have successfully recruited and trained volunteers to assist in translating Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. projects with Cochrane. The type of volunteer and training needed varies across projects. Information about these projects may benefit others involved in translation. RO 20.3 Assessing the impact of bibliographical support on the quality of medical care in patients admitted to an internal medicine service Sarti M1 , Pastori MM2 , Pons M3 , Barazzoni F1 1 Ente Ospedaliero Cantonale (EOC) Head Office, Bellinzona, Switzerland; 2 Università degli Studi di Genova (D.S.Sal.), Dipartimento di Scienze della Salute, Genova, Italy; 3 EOC Department of Internal Medicine, Regional Hospital of Lugano, Switzerland Background: Application of the best current knowledge to decisions in individual patients is the key to evidence-based medicine practice. Research of relevant scientific information through direct consultation of textbooks or journals or online resources is a different way to answer the questions that occur during the visit of the patients. Different studies highlight that the ideal information source must be directly relevant, contain valid information and can be accessed with a minimum amount of work for physicians, but only a few studies address, as objective, the impact of literature assistance on clinical outcome. Objectives: To assess and quantify the impact of literature in diagnostic decisions and treatment of patients admitted to an internal medicine service, using evidence-based medicine methodology. Methods: From November 2012 to February 2013, patients who were hospitalised in the internal medicine service and generated questions about medical care, were randomly assigned to two groups: an intervention group (supported by literature research) and a control group. The information obtained from literature was submitted by email to all the medical team within 12 hours of asking the question. Eight-hundred and sixty-six patients were hospitalised in the analysed period; 201 of these generated questions, and were divided into intervention (n = 101) and control (n = 100) groups. In the intervention group, bibliographical research was possible for 98 subjects. The medical team accepted the results and implemented the research for 90.8% of these subjects (89/98). Statistical analyses were carried out on the intention to treat (ITT) and on the per-protocol populations. Results: Bibliographical research had a significant protective effect on transfer to an intensive care unit (RR = 0.30; 95% CI 0.10 to 0.90; X2 = 5.3, P = 0.02) and hospital readmissions were also influenced by bibliographical research (RR = 0.42; 95% CI, 0.17 to 1.0; X2 = 3.36, P = 0.05) in the ITT population. Conclusions: Our results point out the importance of bibliographical support on the quality of medical care. In particular, they show its possible impact on clinical outcome. Cochrane Database Syst Rev Suppl 1–327 (2015) 131 RO 20.4 The Ptolemy Project: knowledge translation and equity in East Africa Hagstrom C1 , Kendall S2 1 University of Toronto, Canada; 2 Mt. Sinai Hospital, Canada Background: Many practitioners in East Africa rely on Google for literature searches, resulting in untold amounts of irrelevant or non-scholarly material in addition to lack of access to full text. The Ptolemy Project was established in 2001 to allow surgeons in East Africa access to the University of Toronto’s online resources. Participation has increased to include physicians in all medical fields. Objectives: The Project offers a means of obtaining up-to-date information for clinical decision support via the online resources (databases, e-journals, e-books) available from the University of Toronto Library as well as additional resources directly from individual publishers. The University of Toronto library system is ranked third in North America, after Harvard and Yale, with over 1,000,000 electronic resources. Methods: Registration is required; the registration form is on the Project’s website. Registrants are vetted by staff at the Mount Sinai Hospital library in Toronto. Participants log in with a computer-generated username and password and can then link to available resources. Usage is periodically checked; unused accounts are cancelled to allow others to register, as there is a finite number of participants. Results: Registration is selective and participation now stands at about 450, up from 90 in 2001. A brief survey was distributed to all users in July 2014, with positive results. Respondents reported that access to material through Ptolemy improved their practice and patient outcomes. Conclusions: The Ptolemy Project is an invaluable means for physicians and clinical residents in East Africa to obtain both up-to-date and historical health information that would not otherwise be readily available. RO 20.5 Use of research evidence in newborn and child health policies and practices in Kenya Karumbi J1 , Mulaku M2 , Opiyo N3 , English M3 1 Ministry of Health, Kenya, Kenya; 2 School of Pharmacy, University of Nairobi, Kenya; 3 Kemri-Welcome Trust Research Programme, Kenya Background: The process of evidence synthesis began in a very informal process; three topics for which people felt there were gaps were chosen, namely: sickle cell disease, umbilical cord care and fluid therapy in children. Systematic reviews in these areas were conducted and published. Objectives: To develop recommendations for newborn care. Methods: A multi-disciplinary panel was selected jointly by Ministry of Health and Kenya Paediatric Association. The full systematic reviews, short summaries and GRADE (Grading of Recommendations Assessment, Development Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. and Evaluation) ’Summary of findings’ tables were sent to the panel four weeks before the meeting. At the meeting, the panel members were exposed to one day of training on the GRADE process. Evidence was presented for each of the three topics (one topic per day), followed by discussions facilitated by the panelists. The discussions mainly focused on, balance of benefits/harms, feasibility, costs, acceptability (local values), and equity of the proposed interventions. Consensus was developed through voting where necessary. Results: The panels were able to come up with three recommendations that have been incorporated as policies by the Ministry of Health. 1. For sickle cell disease: Hydroxyurea should be considered for use in children under five years of age with a severe form of sickle cell disease where minimum monitoring conditions and appropriate formulation are available; this recommendation was based on low quality evidence. 2. For umbilical cord care: Chlorhexidine 4% should be applied to the umbilical cord immediately after birth, and thereafter daily until the cord separates in babies of > 1000 g (or > 28 weeks) born in facilities; recommendation based on moderate quality evidence. 3. For fluid therapy: In children with severe febrile illness and impaired circulation without signs of severely impaired circulation maintain hydration with appropriate maintenance fluids and do not give a rapid fluid bolus; recommendation based on high quality evidence. Conclusions: The involvement of key stakeholders at the development of recommendations is key and possible even in low-resource settings. RO 20.6 Impact of Cochrane Corners in Portuguese journals Sousa R1 , M Fernandes R2 , Caldeira D3 , Costa J3 , Vaz Carneiro A3 1 Department of Pediatrics, Hospital Santa Maria, Lisbon Academic Medical Centre, Portugal; 2 Department of Pediatrics, Hospital Santa Maria, Lisbon Academic Medical Centre; Portuguese Branch of the Iberoamerican Cochrane Centre (Cochrane Portugal), Lisbon, Portugal, Portugal; 3 Portuguese Branch of the Iberoamerican Cochrane Centre (Cochrane Portugal), Lisbon, Portugal Background: Evidence summaries such as Cochrane Corners are key knowledge translation tools to improve dissemination of high-quality evidence to different end-users. Objectives: To describe and evaluate the impact of Cochrane Corners published in three major Portuguese medical journals by the Portuguese Branch of the Iberoamerican Cochrane Center (Cochrane Portugal). Methods: We developed an editorial partnership with three medical journals published in Portugal: Acta Médica Portuguesa (general medical journal), Acta Pediátrica Portuguesa (pediatrics journal) and Revista Portuguesa de Cardiologia (cardiology). Topics for Cochrane Corners were prioritized based on relevance and novelty. Corners were developed based on previous models, consisting of concise summaries of individual Cochrane Reviews or Overviews, complemented by a pragmatic clinical commentary that puts the evidence in the context of Cochrane Database Syst Rev Suppl 1–327 (2015) 132 national guidelines and practices. We sought to engage an inclusive and diverse set of clinicians authors for each Corner, based on clinical background, training, level of care, and geographical representability; the Branch provided editorial and methodological guidance at all stages. Available metrics of access by Corner and journal include paper downloads and user data (country, platform, link). Results: We will analyze recently published Corners and present metrics of impact for these papers at the Cochrane Colloquium. As measures of relative impact, we will compare these data with access statistics for other published articles in each journal, as well as for individual systematic reviews or overviews that Corners were based on, with a focus on Portuguese-speaking countries. We will also provide a perspective on the challenges and benefits of this approach, particularly regarding the interest of an editorial partnership, the required workload, and the dissemination of Cochrane content for different stakeholders. Conclusions: Cochrane Corners are successful formats to disseminate evidence from Cochrane Reviews. Tailored content that puts the evidence in a local clinical perspective is key to improve its impact. RO 20.7 Best evidence for better practice: using social media to help nurses engage with evidence Chapman S1 1 Cochrane UK, United Kingdom Background: In February 2014, a Cochrane Review was published that found no evidence that the use of structured pressure ulcer risk assessment tools, which is a standard practice in National Health Service (NHS) hospitals in the UK, reduces pressure ulcer incidence. In April, the review was cited in new National Institute for Health and Care Excellence (NICE) guidance making the use of a tool an optional addition to clinical judgement. Objectives: We wanted to share this with the nursing community and encourage them to reflect on their practice in the light of the evidence. Methods: We engaged with nurses though Twitter and our weekly blog, Evidently Cochrane. Other opportunities arose from this initial activity: co-hosting two tweetchats (scheduled, live discussions on Twitter) with the WeNurses community and involvement in Stop The Pressure Month, led by NHS England with the support of The Royal College of Nursing (The RCN). Results: Sharing the evidence on Twitter generated responses from nurses, who were surprised by the review’s findings, and some of these were incorporated into a blog. In March, over 100 participants joined a tweetchat to discuss the evidence. In November, Stop The Pressure Month, a follow-up tweetchat was held, to see if anyone had initiated changes in practice. This, and other recent Cochrane evidence on pressure ulcer prevention and management strategies, was shared in blogs on our own site and on The RCN’s blog site ‘This Is Nursing’. Nurses gave reasons why they would continue to use the tools, for purposes other than that for which they are intended, such as demonstrating activity and strengthening Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. requests for equipment. Conclusions: Social media was used very effectively to disseminate the evidence and engage nurses in debate about research and practice. It enabled us to widen our audience and raise awareness of Cochrane. New relationships with influential individuals and organizations were forged online which resulted in offline meetings and opportunities for future collaboration and communication. RO 20.8 Dissemination of Cochrane Reviews in social media channels and the influence of ‘positive’ versus ‘negative’ results Cuello-Garcia C1 , Perez-Gaxiola G2 , Colunga-Lozano LE3 , Delgado-Figueroa N4 , Agarwal A5 1 Department of Clinical Epidemiology and Biostatistics, McMaster University, Canada; 2 Cochrane Mexico, Hospital Pediatrico de Sinaloa, Mexico; 3 Department of Internal Medicine, Hospital Civil de Guadalajara ‘‘Dr. Juan I. Menchaca’’, Mexico; 4 Department of Evidence-Based Medicine, University of Guadalajara, Mexico; 5 University of Toronto, Canada Background: Currently there is a tendency to measure the research impact of scholarly articles in social media sites (e.g. Twitter, Facebook, Scientific Blogs, mass media, etc.). The Cochrane Library uses Altmetrics, a new tool to encompass social activity in the form of mentions on social media channels, online scholarly activity and commentaries. Objectives: To assess how Cochrane Reviews are disseminated in social media channels and if those with ‘positive’ results (i.e. with a statistically significant result) are more disseminated than those with ‘negative’ results. Methods: We analysed all systematic reviews published in the Cochrane Library from January to December 2013. Metrics obtained from each review included the number of Tweets, Facebook posts, Blog posts, News media, F1000 evaluations, and the total Altmetric score. Reviews were classified as ‘positive’ if the main outcome effect presented a statistically significant result (P value < 0.05, or if confidence intervals reject the null hypothesis) or ‘negative’ if there was not a statistically significant result, or evidence from individual studies was insufficient or null. Results: A total of 1007 reviews were evaluated; 37 had been withdrawn from the Cochrane Library and were excluded from the analyses. The 970 reviews included had a median and interquartile range (IQR) of seven included studies (2 to 15), four Tweets (1 to 11), and an Altmetric score of 4 (1 to 11). Reviews with ‘positive’ results had more tweets than the ‘negative’ reviews (11 [5 to 23] vs 4 [1 to 10]; P value < 0.001) and a higher Altmetric score (6 [2 to 18] vs 3 [1 to 8]; P value < 0.001). After adjusting for number of included studies, country of first author, and status of the review, the differences remained statistically significant. Conclusions: Twitter is the main channel used for disseminating Cochrane Reviews followed by Facebook. Reviews with positive results are tweeted and disseminated in social media channels more often than their negative counterparts. Cochrane Database Syst Rev Suppl 1–327 (2015) 133 RO 20.9 New evidence, new ways to share it, new audiences. Pushing the boundaries with social media Chapman S1 1 Cochrane UK, United Kingdom Background: Many Cochrane reviews relate to common health conditions and could be used by many people making decisions about their health. An obstacle is that they may never have heard of Cochrane, perhaps know nothing about evidence-informed decision-making and aren’t looking for evidence. Social media offers unprecedented opportunities to reach these new audiences. Objectives: Cochrane UK shares evidence primarily though a weekly blog, ’Evidently Cochrane’, and Twitter. We had established a format for the blogs, mostly aimed at patients, which set the evidence in context, gave the main findings and a key message and talked about the strength of the evidence. We sought to expand our audience by experimenting with new ways to present Cochrane evidence, which would make it engaging, relatable and easy to understand. Methods: We introduced quarterly campaign weeks on popular topics (men’s health, palliative care, the menopause), linking in with national events. We invited blogs and commentary from ‘experts’, both health professionals and patients (our youngest blogger being just nine), exploring evidence in the context of experience. Through Advent we published daily posts in varied formats from Lego animations to a film of a pantomime cast talking about Cochrane evidence. We have also blogged responsively to feature evidence relating to health topics that came under the media spotlight. Results: Cochrane UK has a growing following and we are enjoying frequent interactions on Twitter. The blog has been selected by public vote as one of the top ten health organization blogs in the forthcoming UK Blog Awards 2015. Our activities have helped us establish new relationships and created opportunities for collaboration. Conclusions: Innovative use of social media is enabling us to increase our reach, bring evidence to new audiences and create new partnerships. We’re looking forward to seeing what we can do next! Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Rapid oral session 21 Systematic review and meta-analysis methods RO 21.1 How often does an individual trial agree with its corresponding meta-analysis? A meta-epidemiologic study Threapleton D1 , Tam W2 , Di M1 , Tsoi K3 , Tang JL1 1 The Hong Kong Branch of the Chinese Cochrane Centre, The Chinese University of Hong Kong, Hong Kong; 2 Alice Lee Centre for Nursing Studies, Yong Loo Lin School of Medicine, National University of Singapore, Singapore; 3 Jockey Club School of Public Health and Primary Care, The Chinese University of Hong Kong, Hong Kong Background: Evidence from multiple randomized controlled trials (RCTs) accumulates over many years and waits for meta-analysis to provide ‘conclusive’ evidence of treatment effect. This practice may delay the implementation of potentially beneficial treatments. Additionally, even after sufficient evidence has built up, RCTs are still conducted, potentially wasting resources or introducing harm. Objectives: To examine how often the result of the first trial, last trial or any trial selected at random agree with the corresponding meta-analysis. Methods: Published meta-analyses were identified from five key medical journals and also from the Cochrane Database of Systematic Reviews. Meta-analyses were included where an intervention was compared with placebo, no intervention or usual care. The effect size of the earliest, most recent or any trial in the meta-analysis was compared with the meta-analysis effect size. Results: A total of 647 meta-analyses were included; the median number of included studies was 5.2. In 36% (n = 233) of meta-analyses, the earliest published study reported a significant result and the corresponding meta-analysis was significant and in the same direction 84.1% (95% confidence interval (CI) 79.4 to 88.8) of the time. When the earliest study did not have a significant result, the meta-analysis was non-significant in 57.9% (95% CI 53.2 to 62.8) of cases. Similar results were observed when the last or any trial, were substituted in analyses. No instances were identified where the earliest trial and meta-analysis were both significant but in the opposite direction. In meta-analyses with five or more studies, the median time from first to fifth study was 6.5 years. Conclusions: The first trial (or any trial to hand) demonstrating clinical effectiveness is likely to reflect the true treatment effect. This finding has important implications in urgent or critical circumstances where no other effective treatment is available. Cautious recommendations, depending on circumstances or potential harms, may be made if the first trial demonstrates effectiveness. Effective treatment could be implemented many years in advance of a meta-analysis. Cochrane Database Syst Rev Suppl 1–327 (2015) 134 RO 21.2 Confounding, effect modification and the odds ratio: common misinterpretations Shrier I1 , Pang M1 1 Centre for Clinical Epidemiology, Lady Davis Institute, McGill University, Canada Background: When an outcome is dichotomous and investigators are concerned about potential confounding or effect modification across subgroups (e.g. diabetics versus non-diabetics), they often report both the crude (unadjusted) odds ratio and the stratum-specific odds ratios (or adjusted odds ratios). When the stratum-specific odds ratios are different from each other in the absence of bias in either observational studies or randomized controlled trials, authors often interpret this as causal effect modification (biological interaction). Objectives: To illustrate that between stratum-specific odds ratios are actually expected to be different if the variable of interest affects the prevalence of the outcome, even when both causal effect modification and bias are absent. Methods: We demonstrate how and why this phenomenon occurs using hypothetical data from a randomized trial where the one-year untreated mortality was 52%, and the proportion of diabetics (a cause of the outcome) was 30% in each group. Results: In our example, the relative risk was 50% in all patients, in non-diabetic patients, and in diabetic patients. However, the odds ratio was 0.32 in all patients, 0.38 in non-diabetic patients and 0.17 in diabetic patients, with the interaction term from the statistical model being statistically significant. Remembering that the odds ratio becomes more extreme compared to the relative risk as the prevalence increases, our results are a truism since diabetes affects the outcome in our example. The difference between the stratum-specific odds ratios is dependent on the combined effect of the variable that affects the outcome prevalence (proposed effect modifier), and the baseline risk. Conclusions: Although logistic regression is an important tool and reporting adjusted odds ratios (or Cox regression and rate ratios) is appropriate in many contexts, investigators and readers should be wary of claims of effect modification or biological interaction when the covariate is known to be an independent cause of the outcome, and the disease is common. RO 21.3 Cross-over trials –are we doing it right? Are they doing it right? Nolan S1 , Dwan K2 1 University of Liverpool, United Kingdom; 2 Cochrane Cystic Fibrosis and Genetic Disorders Group, United Kingdom Background: Previous work by Elbourne (2002) indicated that the description in the methods section of systematic reviews regarding how cross-over trial data were to be included was insufficient and recommendations were made Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. for improvement. Recent work found that reports of cross-over trials often omit important methodological information relating to design, analysis and presentation of results. Objectives: - To assess review methodology for including cross-over trials in reviews published by the Cochrane Cystic Fibrosis and Genetic Disorder (CFGD) Group and to investigate whether there has been an improvement since the Elbourne results published in 2002. - To assess the quality of reporting of cross-over trials within the trial reports themselves, and subsequently, within Cochrane Reviews. - To establish recommendations for how cross-over data should be included in Cochrane Reviews, in line with the Methodological Expectations of Cochrane Intervention Reviews (MECIR) standards. Methods: We accessed 135 CFGD reviews (published to January 2015) with 655 included studies; 104 reviews indicated cross-over studies were eligible for inclusion and 52 included at least one cross-over study. We identified 212 unique cross-over studies. We checked the methods sections of reviews to identify how review authors planned to manage cross-over data. We accessed cross-over trial reports, the quality of reporting of the results, and compared them to the data included in the review for the primary outcomes. Results: Methods sections of reviews varied in detail. Half of the reviews included specific details regarding paired analyses, Elbourne ‘3 stage’ method or first period analysis, the other half made no statement regarding cross-over trials, but referred to the work by Elbourne without giving further details, intended to consult the Cochrane Handbook or a statistician or intended to analyse cross-over data as parallel data. Full results regarding the quality of reporting within the trial reports and how results were actually included in the reviews will be presented. Conclusions and recommendations at study and review level will also be presented. RO 21.4 How do Cochrane review authors interpret subgroup analyses? Richardson M1 , Garner P1 , Donegan S1 1 Cochrane Infectious Diseases Group, United Kingdom Background: Treatment by covariate interactions can be explored in reviews using subgroup analyses. Such analyses can identify how treatment effect varies by subgroup and are an important methodological approach for stratified medicine. A review of 52 recently published Cochrane Reviews found that 63% applied subgroup analyses [1]. However, interpretation of subgroup analyses in Cochrane Reviews is poor; the review found that only 3% of reviews reported whether or not there was an interaction (i.e. a difference between subgroup results) [1]. Reasons for the lack of interpretation have not been explored. Objective: To identify how review authors interpret subgroup analyses and why analyses are not always interpreted. Methods: We will develop and pre-pilot survey/interview questions based on subgroup analysis scenarios (i.e. no interaction; qualitative and/or statistically significant interaction). We will ask review Cochrane Database Syst Rev Suppl 1–327 (2015) 135 authors: (1) to interpret each analysis; (2) whether there is a statistically significant interaction; (3) how they decided if the interaction was statistically significant. We will also ask why subgroup analyses were not interpreted in their review (if applicable). We will survey correspondence authors for the 52 Cochrane reviews that were included in the recent review [1]. The review included the most recently published review for each Review Group published in the Cochrane Database of Systematic Reviews (searched 8 August 2013). Review authors based locally will be interviewed. Results: We will summarise qualitative results (e.g. interpretations of analyses) and quantitative results (e.g. number of review authors who correctly identified a significant interaction). Review authors will be anonymised. Conclusions: This research will identify how reviews authors interpret subgroup analyses, why subgroup analyses are often not interpreted, and improve how analyses are interpreted in future reviews. Reference 1. Donegan S, Williams L, Dias S, Tudur-Smith C, Welton N. Exploring treatment by covariate interactions using subgroup analysis and meta-regression in Cochrane Reviews: a review of recent practice. Submitted to PLoS one. RO 21.5 Is the risk difference really a more heterogeneous measure? Shrier I1 , Poole C2 , Vanderweele TJ3 1 Centre for Clinical Epidemiology, Lady Davis Institute, McGill University, Canada; 2 Department of Epidemiology, University of North Carolina at Chapel Hill, USA; 3 Department of Epidemiology, Harvard School of Public Health, Harvard University, USA Background: There are claims in the literature that the risk difference is a more heterogeneous measure than the odds ratio or risk ratio. These claims are based on surveys of meta-analyses showing that tests reject the null hypothesis of homogeneity more often for the risk difference than for ratio measures. Objectives: To illustrate that differences in results of homogeneity tests across different scales may be related to statistical power rather than differences in homogeneity itself. Methods: We use hypothetical examples where lack of homogeneity is arguably the same across different scales, but the power (and therefore the expected P value) for the different scales are remarkably different. Results: In the first example, we simulated 75 participants per treatment group where the probability of outcome is 0.206 in Population A regardless of treatment. In Population B, treatment increased the probability of remission from 0.270 to 0.460. Simulating a meta-analysis based on these effects would yield a power of 47% for risk difference, but only 35% for the odds ratio. These numbers exactly match the results used by others to suggest the risk difference is more heterogeneous than the odds ratio. We also simulated 150 participants per group, where baseline remission probability was 0.2 in Population A and B. In this example, treatment increases probability of remission to 0.304 in Population A, and to 0.350 in Population Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. B. but has an effect in Population B, leading to different risk differences and odds ratios in the two populations. These numbers also recreate a power of 47% for risk differences and only 35% for the odds ratios. The effect can also be reversed, with the test of homogeneity having less power for the risk difference compared to the odds ratio. Conclusions: Because current methods cannot be used to conclude that one scale is more heterogeneous than another within meta-analyses of the same data, claims that the risk difference is more heterogeneous should be considered, at best, tentative. The meta-meta-analytic results, therefore, do not support a policy of routinely shunning any of the three measures, including the risk difference. RO 21.6 Small studies are more heterogeneous than large ones: a meta-meta-analysis IntHout J1 , Ioannidis JP2 , Borm GF1 , Goeman JJ1 1 Radboud University Medical Center, The Netherlands; 2 Stanford University, USA Background: Between-study heterogeneity plays an important role in random-effects models for meta-analysis. Most clinical trials are small, and small trials are often associated with larger effect sizes. Possibly there is also a difference in the extent of between-study heterogeneity (τ ) between small and large trials. Objectives: To evaluate empirically whether there is a relationship between trial size and heterogeneity (τ ), using meta-analyses from the Cochrane Database of Systematic Reviews from 2009 to 2013. Methods: We selected the first meta-analysis per intervention review with a dichotomous (n = 2009) or continuous (n = 1254) outcome. The association between estimated τ and trial size was evaluated within meta-analyses using a Bayesian hierarchical model and across meta-analyses using regression. Small trials were predefined as those having standard errors over 0.2 standardized effects. Results: Most meta-analyses were based on few studies (median 4, Q1 to Q3: 2 to 6), and most (74%) primary studies were small. Within the same meta-analysis, the small-study τ S2 was larger than the large-study τ L2: mean ratio 2.11 (95% Credible Interval 1.05 to 3.87) for dichotomous and 3.11 (95% Credible Interval 2.00 to 4.78) for continuous meta-analyses. The imprecision of τ S was larger than of τ L: median SE 0.39 versus 0.20 for small-study and large-study meta-analyses with a dichotomous outcome and 0.22 versus 0.13 for those with a continuous outcome. Similar results were found in the across-meta-analyses approach. Conclusions: Heterogeneity between small studies is larger than between larger studies. The large imprecision with which τ is estimated in a typical small-studies’ meta-analysis is another reason for concern and sensitivity analyses are recommended. Cochrane Database Syst Rev Suppl 1–327 (2015) 136 RO 21.7 RO 21.8 Evaluation of statistical heterogeneity measures and their precision in systematic reviews with meta-analysis and the impact in their interpretation Systematic reviews should explore a priori hypotheses to explain heterogeneity even when I2 is low Ruiz Gaviria RE1 , Rodriguez Malagon MN1 1 Clinical Epidemiology and Biostatistics Department, Faculty of Medicine, Pontificia Universidad Javeriana Medical School, Bogota, Colombia Background: Systematic review, followed by meta-analysis, is one of the most powerful and widespread tools for understanding efficacy and safety evaluation of medical interventions. However, statistical heterogeneity is one of the most serious issues when conducting this type of analysis. Cochran Q coefficient and I2 (I-squared) are the main statistics used to evaluate it. Confidence intervals (CI) can be calculated for both of them in order to determine their precision. Objectives: The aim of this study was to review the application of the method to determine statistical heterogeneity, and its precision, in a group of selected systematic reviews that also considered meta-analysis and are published by Cochrane. Methods: We searched in PubMed and Wiley for systematic reviews with meta-analysis that investigated chronic obstructive pulmonary disease and were produced by the Airways Group and published on the Cochrane Library. We included systematic reviews published between January 2013 and February 2015. We extracted the number of studies used in subgroup and total analysis, the Q coefficient, its P value, I2 statistics and their CI. If the CI was not present, we calculated it using the method proposed by Higgins et al. Results: We included five reviews in the analysis. None reported the CI of the I2. The calculated CIs were consistently large, showing heterogeneity that could be interpreted as mild to severe. We found inconsistencies between the interpretation of the heterogeneity of the studies and the calculated CI in all of them, especially when the authors stated that there was no heterogeneity. Conclusions: Heterogeneity is a serious concern in meta-analysis. We found inconsistencies between the interpretation of the reported values and the CIs of the I2. We propose that authors and also reviewers consider the existence of heterogeneity as an issue that really affects meta-analysis results because the selection of whether to apply a random-effects or fixed-effect model depends on its presence. Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Siemieniuk R1 , Meade M1 , Alonso P2 , Briel M1 , Vandvik P3 , Guyatt G1 1 McMaster University, Canada; 2 Iberoamerican Cochrane Center, Biomedical Research Institute (IIB-Sant Pau-CIBERESP), Barcelona, Spain; 3 Norwegian Knowledge Centre for the Health Services, Oslo, Norway Background: There is general agreement that systematic review authors should generate a priori hypotheses to explain heterogeneity and test these hypotheses when heterogeneity proves substantial. However, when the meta-analysis suggests low heterogeneity, as represented by a low I2, controversy exists. In these circumstances, some advocate, and practice, omitting statistical exploration of heterogeneity. Others disagree. Objective: To illustrate the advisability of exploring possible subgroup effects even when I2 is low. Method: We conducted a systematic review and meta-analysis addressing the desirability of adjunctive administration of corticosteroids in patients with community-acquired pneumonia. We generated four a priori hypotheses to explain heterogeneity, including the severity of pneumonia (expectation of larger effect on mortality when over 70% of patients had severe pneumonia). Results: Random-effects meta-analysis showed a relative risk (RR) of 0.67, 95% confidence interval (CI) 0.47 to 0.97, I2 =7%, for overall mortality (Figure 1). Despite the low I2 we undertook Chi-square tests for effect modification for our a priori hypotheses. We found an apparent mortality benefit in trials that met our ’more severe’ criteria (6 studies; n = 388; RR = 0.39, 95% CI 0.22 to 0.67; I2 = 0%) but not in those that did not (6 studies; n = 1586; RR = 1.00, 95% CI 0.64 to 1.56; I2 = 0%; interaction P = 0.009; Figure 2). The subgroup finding gains credibility from the large magnitude of effect, its biological plausibility (a greater inflammatory response in more severe pneumonia), the small number of a priori hypotheses with specified direction, and a small interaction P value. It is based, however, on differences between studies rather than within studies, and was driven to a considerable extent by a small study that was stopped early for benefit and almost certainly represents a large overestimate of effect. Overall, the credibility of the subgroup effect is moderate. Conclusions: A low I2 should not deter systematic review authors from exploring a priori hypotheses to explain heterogeneity. Cochrane Database Syst Rev Suppl 1–327 (2015) 137 Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Cochrane Database Syst Rev Suppl 1–327 (2015) 138 RO 21.9 Adding non-randomized and non-controlled studies to an existing review on the effectiveness of chronic disease management programs for asthmatic patients Arditi C1 , Burnand B1 , Peytremann-Bridevaux I1 1 Cochrane Switzerland and Institute of Social and Preventive Medicine, Lausanne University Hospital, Switzerland Background: Systematic reviews (SRs) of interventions usually include randomized controlled trials (RCTs) only. However, complex interventions such as chronic disease management (CDM) programs are often evaluated using study designs other than RCTs, mainly for reasons of feasibility. Considering non-randomized and non-controlled (NRNC) studies in an SR on CDM programs could be informative both in terms of intervention context and effectiveness. Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Objectives: To assess if and how setting, population, intervention characteristics, and results varied when adding NRNC studies to an ongoing Cochrane SR with the Effectivea Practice and Organization of Care Review Group. Methods: All NRNC studies identified as evaluating CDM programs in asthmatic adults, but excluded from the Cochrane SR because of the design, were considered in this review, in addition to the already included studies. Two review authors independently extracted data and assessed the methodological quality using the Cochrane risk of bias tool, with additional questions for non-controlled studies. We compared study and intervention characteristics, as well as effectiveness, by study design. Results: Thirty-seven studies were added to the 20 studies already included in the original SR. The added studies were of poor methodological quality (e.g. no statistical adjustments for confounding in controlled before-after studies). Study and intervention characteristics of the NRNC studies differed from those in the original SR (Table 1). For instance, CDM programs were implemented in a larger variety of countries, Cochrane Database Syst Rev Suppl 1–327 (2015) 139 included larger numbers of patients, and reported process and healthcare use outcomes more often. Effectiveness also varied according to study design (Table 2). Overall, the proportion of outcomes that improved in response to CDM was higher in the NRNC studies compared with the studies from the original SR. Conclusions: While NRNC studies provide results from interventions in real-life settings that may be valuable to policymakers and other stakeholders implementing CDM programs, the methodological quality of the studies is poor, limiting the validity and the interpretation of their results. Rapid oral session 22 Guidelines and GRADE RO 22.1 Strength of the evidence and treatment recommendations: a survey of the Cochrane Database of Systematic Reviews for the first listed primary outcome in SoFs in each review and reasons for upgrade or downgrade were recorded. For reviews with high quality of evidence, those with low quality of evidence, and a random sample of 50 reviews with no GRADE assessment, we also extracted the results for the evaluated outcomes and recommendations for the use of interventions. Results: A total of 1394 SRs were identified; 608 (43.6%) of these incorporated GRADE. Within these reviews 13.5%, 30.8%, 31.7% and 24.0% reported high, moderate, low and very low levels of evidence, respectively, for the evaluated primary outcome. High level quality was more common in updated than new reviews and for pharmacological than other types of interventions. Even when all outcomes listed in the SoFs were considered, only 116/608 (19.1%) SRs had at least one outcome with a high quality of evidence. Twenty-five (30.5%) of 82 SRs with high quality evidence had both significant results and interventions recommended by the authors. Conclusions: A minority of systematic reviews in the CDSR have high level evidence and even in this small set, an intervention was recommended in only one-third of the reviews. Fleming P1 , Koletsi D2 , Ioannidis J3 , Pandis N4 1 Barts and the London School of Medicine and Dentistry, United Kingdom; 2 University of Athens, Greece; 3 Stanford University School of Medicine, USA; 4 University of Bern, Switzerland RO 22.2 Background: The GRADE (Grading of Recommendations Assessment, Development, and Evaluation) tool allows systematic appraisal of the quality of a body of evidence based on specific domains leading to an overall score assessment presented in ’Summary of findings’ tables (SoFs). Objectives: To survey the level of evidence contributing to recommendations derived from Cochrane Reviews and how this level of evidence correlates with whether an intervention is recommended or not. Methods: All systematic reviews on the Cochrane Database of Systematic Reviews (CDSR) from 1 January 2013 to 30 June 2014 were identified and selected by two authors from the Cochrane Library. Data were extracted independently by two authors and the quality of the evidence Selph S1 , Holmes R1 , McDonagh M1 1 Pacific Northwest Evidence-Based Practice Center, USA Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Predictors of change in updated quality of evidence ratings in Cochrane Reviews Background: Using GRADE (Grading of Recommendations Assessment, Development and Evaluation tool) to rate the quality of the body of evidence is complex and subjective, and how consistently reviewers apply the rating system is unclear. Objectives: The aim of this study is to evaluate how changes in evidence contribute to changes in quality of evidence ratings by comparing Cochrane Reviews having quality of evidence tables with earlier and later versions. Methods: We conducted a MEDLINE search of the Cochrane Library for systematic reviews with quality of evidence tables Cochrane Database Syst Rev Suppl 1–327 (2015) 140 in sequential updates. For each outcome with a change in number of studies, sample size, or quality rating, we abstracted publication years, review authors, effect estimates, numbers of included studies, and sample sizes and conducted regression analyses to determine predictors of a change in quality rating. Results: A total of 1814 systematic reviews were marked in the Cochrane Library as recently updated. Preliminary analysis of 350 reviews found nine that provided quality ratings tables for sequential updates, representing 36 different outcomes. Two factors associated with a change in quality rating in univariate analysis were: change in numbers of participants (P value 0.041) and change in magnitude of the effect estimates (P value 0.008). Changes in precision or statistical significance of effect estimates were not associated with a change in quality rating. Quality of the evidence was downgraded (n = 13) more often than upgraded (n = 4) and was sometimes downgraded with no change in the evidence. Conclusions: Changes in the magnitude of effect estimates and numbers of participants were associated with changes in quality of the body of evidence ratings. Quality ratings were downgraded more often than upgraded which may indicate a reviewer ’experience’ factor as the reasons for downgrading were not always obvious and it would be helpful if reviewers explained the rationale for these changes. Further analysis including the remainder of the reviews and the use of multivariable regression to explore relationships between variables is in progress. RO 22.3 Reported use and perceptions of value of Cochrane evidence by South African guideline developers 1 1 2 3 1 4 Abrams A , Kredo T , Young T , Louw Q , Grimmer K , Daniels K 1 South African Cochrane Centre, South African Medical Research Council, South Africa; 2 South African Cochrane Centre, South African Medical Research Council and Stellenbosch Univeristy, South Africa; 3 Stellenbosch University, South Africa; 4 South African Medical Research Council, South Africa Background: Systematic reviews are a key source of transparent evidence assessments for guideline development. This research, as part of the South African Guidelines Excellence (SAGE) project, maps the use and perceived benefits of Cochrane Systematic Reviews amongst primary care guideline developers in South Africa. Methods: A modified snow-ball technique identified key stakeholders. Individual interviews were recorded and transcribed. Content and descriptive analysis was undertaken. Interview scripts were analysed using an inductive approach. Results: Twenty-four interviews have been conducted with 23 key stakeholders actively involved with guidelines in government, public and private health care. Early exploratory analysis indicates that guideline development is a complex web of interactions between stakeholders; a process that is informed by individual values, politics and power at both the individual and group level. Initially Cochrane-related work and research Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. was not a focus of this study, but a trend emerged within the interviews. Without prompting from interviewers, 8/22 respondents named Cochrane as a key resource for guideline evidence assessments, or as a key player in reviews of evidence. ‘‘Obviously I’m going to look at Cochrane . . .I’m going to try look at the highest level first. So the issue is what you want, you want integrity . . .’’ (INT01/398). Of the 14 interviewees who did not name Cochrane, four rely on World Health Organization (WHO) guidelines to inform the evidence they use in local guideline development. The remainder describe the need for an evidence grading process, but highlight a variety of barriers to establishing this process in the local context including limited human capacity and time. Conclusion: Within the complex guideline development process in South Africa, it is clear that Cochrane evidence plays a role. Understanding this role can provide Cochrane reviewers with input on key areas of interest for policy and guideline writers, identifying where Cochrane Reviews can cater specifically to the needs of guideline writers, especially in LMICs. RO 22.4 An algorithm to assign GRADE levels of evidence to comparisons within systematic reviews Pollock A1 , Farmer S1 , Brady M1 , Langhorne P2 , Mead G3 , Mehrholz J4 , van Wijck F1 , Wiffen P5 1 Glasgow Caledonian University, United Kingdom; 2 University of Glasgow, United Kingdom; 3 University of Edinburgh, United Kingdom; 4 Klinik Bavaria in Kreischa GmbH, Germany; 5 University of Oxford, United Kingdom Background: An essential part of a Cochrane overview is the assessment of the quality of evidence within included reviews. We planned to use the GRADE (Grading of Recommendations, Assessment, Development and Evaluation) approach in our overview (as recommended in the Cochrane Handbook), but its subjectivity led to inconsistency of application. Objectives: To develop and use an algorithm to assign GRADE levels of evidence objectively to comparisons within reviews included in a Cochrane Overview. Methods: After initial exploration of applying GRADE levels of evidence, authors agreed criteria perceived to be most relevant for judging quality of the particular evidence synthesised within the overview. Key criteria judged to be of most importance were: number of participants; risk of bias of trials, heterogeneity; and methodological quality of the review. An initial algorithm was drafted, applied to a convenience sample of 43 comparisons, and compared to previous, independently-applied, subjective judgements of GRADE level. An iterative process explored impact of criteria ’weighting’ within the algorithm and the number of consequent downgrades of quality level. We created and applied four different formulae to assign downgrades to each of the 43 sample comparisons and explored the resulting levels of evidence. Results: An algorithm, judged to assign the most appropriate levels of evidence (see Table 1), and a formula, for assigning GRADE Cochrane Database Syst Rev Suppl 1–327 (2015) 141 level of evidence based on number of downgrades determined using the algorithm (see Table 2), were developed. This algorithm was applied to all 127 comparisons included within the overview. Conclusions: This algorithm enabled us to justify and report the GRADE level of evidence clearly for each comparison, and also to be consistent, transparent and efficient in decisions around levels of evidence for each comparison. Whilst mechanistic in application, this algorithm aims to capture what was subjectively judged to be of greatest importance to the quality of this specific evidence-base. We propose that this methodological approach has implications for assessment of quality of evidence within future evidence syntheses. RO 22.5 Development of a useful system for grading of recommendations and its application on the Korean Guideline Jang J1 , Shin E1 , Yeon J1 , Chang S2 , Kim D3 , Lee Y4 1 Korean Academy of Medical Sciences, South Korea; 2 Department of Urology, Kyung Hee University School of Medicine, South Korea; 3 Department of Diagnostic Radiology, Yon Sei University School of Medicine, South Korea; 4 Department of Legal Medicine, Seoul National University College of Medicine, South Korea Background: To alleviate discrepancies between evidence and practice, the Korean Modified Grading of Recommendations Assessment, Development and Evaluation (GRADE) system has been developed and applied to clinical practice guideline on diabetes mellitus. Objectives: To review Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Cochrane Database Syst Rev Suppl 1–327 (2015) 142 the determinants for grading system used in Korean clinical practice guidelines (CPGs) and identify the strong recommendations with low quality evidence, and apply the Korean Modified GRADE system. Methods: To review the grading factors of recommendations in Korean CPGs, 43 guidelines from the Korean Medical Guideline Information Center were classified according to a variety of categories. Based on the classification, the Korean Modified GRADE system is developed by reflecting the opinion of end-users like clinical physicians. The Korean Modified GRADE system is then applied to 2014 Evidence-based recommendations for type 2 diabetes in primary care developed with the Korean Academy of Medical Science and the Korean Centers for Disease Control & Prevention, and an online survey was conducted by 30 doctors through the Delphi process. Results: The determinants to be considered in 43 Korean guidelines were ‘balance of benefits and harms’, ‘patient’s value and preference’, ‘quality of evidence’, ‘cost’, ‘expert opinion’ and ‘health setting’. Strong recommendations with low evidence constituted 47 of 215 recommendations, accounting for about 21.9%. After applying the Korean Modified GRADE system, about 12.6% of the total recommendations of ‘2014 Evidence-based recommendations for type 2 diabetes in primary care’ have been adjusted to up-grading. Conclusions: The Korean Modified GRADE system can be utilized as an effective tool to better reflect the opinion of the physicians at medical field. RO 22.6 Do Cochrane Systematic Reviews inform clinical practice guidelines in urology? Gudeloglu A1 , Ozdemir B1 , Dahm P2 1 Prostatic Diseases & Urological Cancers Group, Turkey; 2 Prostatic Diseases & Urological Cancers Group, USA Background: Clinical practice guidelines (CPGs) aim to provide evidence-based guidance to inform individual patient and health policy decision-making. They should ideally be based on high quality systematic reviews, for example as provided by Cochrane. Uptake of Cochrane Reviews by guideline developers also represents an important measure of relevance and impact for Cochrane. Objectives: We performed this study to assess the uptake of Cochrane Reviews in urological guidelines. Methods: We retrieved the most up-to-date version of CPGs by five major guideline developers relevant to urology, namely those of the European Association of Urology (EAU), the American Urological Association (AUA), the National Institute for Health and Care Excellence (NICE), the National Comprehensive Cancer Network (NCCN), and United States Preventive Services Task Force (USPSTF). Two independent reviewers abstracted the number of citations of Cochrane Reviews and the review group from which they came. Results: We identified a total of 66 CPGs; the largest number was published by the EAU (n = 26) and AUA (n = 20), followed by the USPSTF (n = 8), NCCN (n = 6) and NICE (n = 6). Less than half (30/66; 45.5%) of all CPGs cited at least one Cochrane Review. Overall, Cochrane Reviews were cited 199 Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. times. The proportion of CPGs by different organizations that cited at least one Cochrane Review was 83.3%, 61.5%, 50.0%, 20.0% and 16.6% for NICE, EAU, USPSTF, AUA and NCCN, respectively. The largest contributions of cited Cochrane Reviews originated from the Incontinence (29.1%), Prostatic Diseases and Urological Cancers (23.6%), Pain Palliative and Supportive (17.6%), and Renal (12.1%) Groups. Conclusions: The uptake of Cochrane Reviews by CPGs developers for the field of urology is suboptimal, especially in the USA. Increased efforts to promote high quality systematic reviews by the relevant Cochrane groups are therefore critically important. RO 22.7 The predictive validity of GRADE for the stability of effect estimates was low: a meta-epidemiological study Evans T1 , Gartlehner G2 , Dobrescu A3 , Bann C1 , Lohr K1 1 RTI International, USA; 2 RTI International, Cochrane Austria, Danube University, USA, Austria; 3 Victor Babes University of Medicine and Pharmacy, Timisoara, Romania Background: Many organizations have adopted the GRADE (Grading of Recommendations Assessment, Development and Evaluation) approach to rate researcher confidence in an available body of evidence. GRADE’s definition of quality of evidence (QoE) links individual grades to the degree of confidence that estimates are close to the true effect (and thus will remain stable as new evidence accrues). Objectives: To determine the predictive validity of the GRADE approach (i.e. whether GRADE discriminates reliably between evidence that remains stable and evidence that changes as new studies emerge). Research Design/Methods: To determine the predictive validity, we randomly assigned 13 researchers who are producers and users of systematic reviews using the GRADE approach from six US Evidence-based Practice Centers and Cochrane Austria 160 bodies of evidence to grade. Using likelihoods from the survey as reference points, we calculated c-statistics to determine the predictive validity. Results: GRADE did not discriminate well between bodies of evidence that remained stable and those that changed (c-scores 0.56 to 0.58). Conclusions: GRADE is a suitable method for systematic review producers to convey uncertainties to users. The predictive validity of GRADE was compromised by grades of QoE that seemed, in general, too low. Cochrane Database Syst Rev Suppl 1–327 (2015) 143 RO 22.8 Reproducibility of Grading of Recommendations Assessment, Development and Evaluation (GRADE) factors on the strength of recommendations: an empirical assessment Kumar A1 , Miladinovic B1 , Guyatt GH2 , Schunemann HJ2 , Djulbegovic B1 1 Morsani College of Medicine, USF Health, Program for CER, Tampa, USA; 2 McMasters University, Hamilton, Canada Background: GRADE is a widely used methodology for the development of clinical practice guideline but its reproducibility has not been tested in context of development of clinical practice guidelines. Objective: Assess the reproducibility of all GRADE factors among guidelines panel members with limited exposure to GRADE methodology. Methods: The study was conducted as part of the clinical practice guideline development process of American Association of Blood Banking (AABB) for the use of prophylactic versus therapeutic platelet transfusion in patients with thrombocytopenia. The results from systematic review and meta-analysis for each question were summarized as a GRADE evidence profile. Inter-rater agreement for all GRADE factors and strength of recommendations was summarized using a weighted kappa statistic with 95% confidence intervals (CI). Results: Eighteen panel members participated in the deliberation of making recommendations and completed the online questionnaire. They were given two one-hour lectures about GRADE. The inter-rater agreement for the domain of quality of evidence was good (kappa value: 0.68; 95% CI 0.541 to 0.837), and fair for balance of benefit and harms (kappa value: 0.4; 95% CI 0.253 to 0.574) and use of resources (kappa value: 0.275: 95% CI 0.116 to 0.421). The inter-rater agreement was moderate for the GRADE domain of patients’ values and preferences (kappa value: 0.441; 95% CI 0.307 to 0.555). The inter-rater agreement for making a for/against recommendation was good (kappa value: 0.738; 95% CI 0.331 to 0.914) and fair for strong/weak recommendation (kappa value: 0.391; 95% CI 0.175 to 0.681). Conclusions: While not all elements of GRADE system had Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. good agreement, the inter-rater agreement for assessing the quality of evidence and issuing a recommendation of for, versus against, was substantial. This is probably because GRADE has operationalized these two areas in more details than other domains. Further operationalization of all GRADE domains would likely improve its reproducibility across the entire GRADE system. Rapid oral session 23 Information retrieval RO 23.1 Searching for trials in systematic reviews versus a specialist trials register: a case study comparison of source, time and costs Brunton G1 , Stokes G1 , Shemilt I2 , Stansfield C1 , Sutcliffe K1 , Thomas J1 1 EPPI-Centre, UCL IOE London, United Kingdom; 2 Behaviour and Health Research Unit, Dept. of Public Health & Primary Care, University of Cambridge, United Kingdom Background: Some public health systematic reviews examine interventions that reflect transdisciplinary health concepts, creating challenges in identifying either too many or too few eligible studies. Two sources may be of use: existing reviews’ evidence tables and reference lists, and a specialist public health trials register (TRoPHI). The costs and benefits of searching these sources to identify eligible trials are unknown. Objective: To assess the incremental costs and benefits of searching for eligible trials in evidence tables and reference lists of existing systematic reviews, compared with a TRoPHI search, in a case study systematic review of community engagement interventions. Methods: We searched six multidisciplinary databases for existing reviews, and TRoPHI, to capture records of eligible trials. Two researchers independently screened systematic reviews and TRoPHI records, followed by full-texts, to identify eligible trials. The researchers prospectively tracked the flow of records and studies through screening and time taken to complete the process (searches, screening and duplicate removal). Results: The TRoPHI search retrieved 865 unique records, of which 226 corresponding full-texts were screened to identify 153 reports of eligible trials: the search, screening and duplicate removal process took 31.0 hours. The systematic reviews search retrieved 342 reviews containing 703 unique records of trials, of which 153 were reports of eligible trials: the search, screening and duplicate removal process took 128.5 hours. No additional eligible trial reports were identified by searching existing systematic reviews. Analysis of costs is in process. Conclusions: Searching existing systematic reviews increased the cost of study identification with no incremental benefit to the review in this case. Researchers undertaking complex reviews are encouraged to consider the transdisciplinary nature of their review topic to help determine Cochrane Database Syst Rev Suppl 1–327 (2015) 144 the relative trade-offs for each source in terms of time, effort and yield of unique citations. Further case studies could help establish appropriate methods of searching systematic reviews to source primary studies. RO 23.2 Searching trials registers to identify relevant studies for Cochrane Reviews: verifying the optimal search approaches Metzendorf MI1 , Richter B1 1 Metabolic and Endocrine Disorders Group, Germany Background: The Methodological Expectations of Cochrane Intervention Reviews (MECIR) require Cochrane authors to search ClinicalTrials.gov (CT.gov) and the International Clinical Trials Registry Platform (ICTRP) in addition to bibliographic databases. The evidence on how to search trials registers soundly is evolving. According to previous research, high retrieval performance (defined as best sensitivity) is achieved only when using sensitive search strategies in specific interfaces of the trials registers. Objectives: To determine which search approach in which interface of CT.gov and ICTRP yields the best retrieval performance to identify relevant trials on a review’s topic and to verify previously conducted research. Methods: We tested searches from CMED reviews using the methodology by Glanville 2014. We tested four search strategies per review topic: sensitive (various terms for intervention/condition) and highly precise (single term for intervention/condition) each run in the basic and advanced interface of the trials registers respectively. We did not evaluate the highly sensitive (various terms for intervention) nor the precise search strategy (single term for intervention), as searches developed for most Cochrane Reviews will include intervention combined with condition. Results: The reviews comprised pharmacological, complementary, surgical and nutritional interventions. The conditions included type 2 diabetes mellitus, obesity, idiopathic short stature and cardiovascular disease. In most cases the sensitive approach performed better in both trials registers. In CT.gov both search interfaces performed well, but the advanced search offered an improved precision without compromising sensitivity. In ICTRP the standard search was clearly superior to the advanced search. Conclusions: Our results are consistent with previous findings. When searching trials registers researchers should apply sensitive search strategies including synonyms. In CT.gov both basic and advanced search interfaces can be used for optimal retrieval. In the ICTRP the advance search performes generally poor, therefore researchers should only use the standard search interface. Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. RO 23.3 Efficiently searching for exhaustiveness: how to perform high quality searches faster Bramer WM1 , de Jonge G1 , Mast F1 1 Erasmus MC, The Netherlands Background: Searching exhaustively, as is necessary for systematic reviews, can be very time consuming for both librarians and researchers. Selecting the search terms (balancing sensitivity and specificity) and translating queries between databases is challenging. Objectives: A method was developed for translating a research question into optimal queries in several databases, but are the results good enough, and what are the success factors? Methods: In order to guarantee term completeness an optimization technique was developed. This method easily identifies missed relevant terms. Macros in MS Word have been developed to convert syntax automatically between databases and interfaces. Information specialists have used these methods to create searches for systematic reviews. Many parameters of the search are recorded in the process, and after completion of the review, the included references are checked for database coverage and retrieval. Results: Over 425 searches have been tracked. For more than 30 of these the development of the optimization process, and for more than 100 searches, the relevant articles have been chosen by the reviewers. The parameters and outcomes of these searches is compared to data from published reviews from other Dutch University Medical Libraries. Ninety per cent of the searches performed with this method by one of the authors were completed within two hours. The median time needed was 70 minutes. Still, the searches are longer than the comparative data, use more databases, retrieve more relevant references, and have a higher precision. Conclusions: The information retrieval process can be shortened, without loss of quality. This allows us to assist many more review projects than was possible with conventional methods, increasing the overall quality of our institutes’ scientific research. RO 23.4 Is it possible to focus EMTREE without loss of sensitivity when searching Embase for systematic reviews? Evidence from practice Duffy S1 , Ross J1 , Misso K1 , Noake C1 , Stirk L1 1 Kleijnen Systematic Reviews Ltd, United Kingdom Background: Systematic reviews require a comprehensive search of numerous databases in order to minimise bias. MEDLINE and Embase are the most commonly searched bibliographic databases when undertaking reviews of healthcare interventions. As the overall search results for systematic reviews appear to be getting increasingly larger, it would help reduce workload if search results could be made smaller. Focusing EMTREE subject heading terms in Embase could significantly reduce the number of records retrieved. Cochrane Database Syst Rev Suppl 1–327 (2015) 145 Objectives: To investigate whether restricting EMTREE indexing terms in order to focus when searching Embase will reduce the number of records retrieved without loss of relevant studies. Methods: Embase searches conducted in recent systematic reviews undertaken by Kleijnen Systematic Reviews will be retrospectively compared with search strategies in which the EMTREE terms have been focused. The records retrieved by the focused EMTREE search will be investigated to see if included studies identified by the original unrestricted Embase search strategy have been retrieved. The review searches under investigation will cover various subjects, such as anaphylaxis, bile acid malabsorption, and KRAS for colorectal cancer. A prospective investigation will be undertaken on a current review: medicinal cannabis. Results: The data collected will be analysed to identify: total with and without restriction to focus, yield of included records, and Number Needed to Read (NNR) to detect relevant references. Conclusions: The project will explore overall yield and recall of relevant included records by each approach, and whether focussing EMTREE terms can reduce screening burden without significantly impairing recall of relevant records. Reducing the number of records retrieved from systematic review searches without a loss of sensitivity will improve efficiency, save time and minimise costs. Only a few, however, were included in the review, because most of authors did not search the Ichushi-Web database of the Japan Medical Abstracts Society (JMAS). Objectives: We examine whether results from the existing reviews may be influenced by those of the Ichushi-Web, when articles did not appear in PubMed. Methods: Using the literature search strategy in the review by Chartrand 2012, we searched for Japanese and English studies on the Ichushi-web between 2000 and March 2015. We extracted outcomes regarding RIDTs accuracy and assessed risk of bias. Results: From the Ichushi-Web, four studies written in Japanese matched the inclusion criteria for RIDTs accuracy for adults and children. These are shown in Table 1. Conclusions: JMAS was established in 1903, and more recently launched Ichushi-Web, which was an exhaustive collection of Japanese biomedical literature for the previous 110 years. The number of citations in the database increases by more than 300,000 per year and now exceeds nine million citations. Results of the RIDTs accuracy review might be influenced by Japanese studies, as four studies seemed to be eligible for the RIDTs accuracy review. Adding a search of a Japanese literature retrieval database might be recommended, especially for certain topics considered in Japanese research studies. RO 23.6 RO 23.5 How do the search results of the Ichushi-Web influence systematic review? The case of rapid influenza diagnostic tests (RIDTs) Kojimahara N1 , Kawai F2 , Morizane T3 1 Tokyo Women’s Medical University, Japan; 2 St. Luke’s International University Library, Japan; 3 Medical Information Network Distribution Service (MINDS), MINDS Guideline Center, Japan Background: Some clinical studies regarding diagnosis for influenza were carried out by primary care physicians and published only in Japanese. Medical settings in Japan tend to have higher rates of adopting rapid influenza diagnostic tests (RIDTs) for suspicious influenza-like illness in outpatients. Searching Chinese biomedical databases: current practice among Cochrane reviewers Cohen JF1 , Korevaar DA1 , Wang J1 , Spijker R2 , Bossuyt PM1 1 Department of Clinical Epidemiology, Biostatistics and Bioinformatics, Academic Medical Center, University of Amsterdam, The Netherlands; 2 Dutch Cochrane Centre, University Medical Center, Utrecht, The Netherlands Background: Cochrane recommends searching the literature without language restrictions to avoid language bias. China is currently ranked second in the world, after the USA, in terms of the total number of published scientific papers. Objectives: To describe the use that Cochrane reviewers make of Chinese biomedical databases. Methods: In a systematic overview of the Cochrane Library from inception to 2014, we assessed Table 1 Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Cochrane Database Syst Rev Suppl 1–327 (2015) 146 the extent to which Cochrane reviewers had searched five major Chinese databases (China National Knowledge Infrastructure (CNKI), Chinese Biomedical Literature (CBM), Chinese Medical Current Content (CMCC), VIP, and WANFANG (China Online Journals)). These Chinese databases index about 2500 journals, of which less than 6% are also indexed in MEDLINE. Results: Only 243 of 8680 published Cochrane Systematic Reviews (3%) searched one or more of the five major Chinese databases. All 243 systematic reviews evaluated an intervention; 179 (74%) had at least one co-author with a Chinese affiliation; 118 (49%) addressed a topic in complementary and alternative medicine, such as acupuncture. Conclusions: Cochrane reviewers rarely include Chinese databases in their literature searches. Further research should evaluate the consequences of this failure to search Chinese databases when performing systematic reviews. RO 23.7 Supplementary searches of PubMed to improve currency of MEDLINE and MEDLINE In-Process searches via Ovid: evidence from practice Duffy S1 , Misso K1 , Noake C1 , Ross J1 , Stirk L1 1 Kleijnen Systematic Reviews Ltd, United Kingdom Background: When conducting literature searches for systematic reviews one bibliographic database, MEDLINE, is almost always included. MEDLINE content can be searched via numerous different search interfaces, the majority of which are provided by fee-based subscription services (e.g. OvidSP), although it can be accessed for free using PubMed. Many information specialists would choose to search using sophisticated interfaces (usually fee-based) that enable the design of complex search strategies. PubMed has more limited search capabilities, for example proximity searching is not possible. However, 2% of PubMed records are not found in MEDLINE, including newly published and ahead-of-print articles. Objectives: To investigate whether conducting a supplementary search of PubMed in addition to the main MEDLINE (OvidSP) search is worthwhile. We will explore whether an extra PubMed search can be conducted quickly and retrieve unique studies that are included in the review. Methods: Searches of PubMed will be conducted after MEDLINE and MEDLINE In-Process (OvidSP) searches have been completed. Due to different levels of functionality in search interface, exact translation of the MEDLINE strategy may not be possible; in this instance the objective is not to duplicate searches, rather to supplement them. Search strategies will be simplified, using phrase searching predominantly. The searches will be limited to records not in MEDLINE, using the following limit: (pubstatusaheadofprint OR publisher[sb] OR pubmednotmedline[sb]) that PubMed searches are as current as possible can determine the timeliness of the final review. It may also be worth establishing search alerts in PubMed until final publication. RO 23.8 The Cochrane EMBASE project for populating the Cochrane Central Register of Controlled Trials (CENTRAL): can we stop searching EMBASE separately? Faulkner S1 , Noel-Storr A2 , Eisinga A3 1 Cochrane ENT, United Kingdom; 2 Cochrane Dementia and Cognitive Improvement, United Kingdom; 3 Cochrane UK, United Kingdom Background: The Cochrane EMBASE project aims to identify randomized and quasi-randomized trials (RCTs/q-RCTs) using highly sensitive, validated search strategies developed by information specialists that are run in EMBASE. The results are screened for eligibility and inclusion in CENTRAL by volunteers using a crowdsourcing model (http://www.cochranelibrary.com/help/central-creationdetails.html). One aim of this centralized search effort is to improve Cochrane Review Group’s Trials Search Co-ordinators’ (TSCs) efficiency by reducing duplication of effort in identifying studies in EMBASE for their registers and for CENTRAL. Objectives: To determine whether the Cochrane ENT Group searches in EMBASE, designed to populate their register and CENTRAL, identify any unique RCTs/q-RCTs not found by the centralised EMBASE search. Methods: The search strategies used are published in the Cochrane ENT module of the Cochrane Library (http://ow.ly/KtDMN). We will run these searches from May 2014 to May 2015 in Ovid EMBASE. We expect to retrieve approximately 3500 references. We will also run the ENT register searches in PubMed and CENTRAL, and upload results to the ENT segment of the Cochrane Register of Studies (CRS) first. We will then upload the records from EMBASE into CRS. We will use the CRS deduplication tool to exclude records already identified as relevant, or already excluded as ineligible. We will then screen the remaining references to identify any RCTs/q-RCTs in ENT that have not been identified through the centralised EMBASE project, the centralised PubMed search or the ENT Group’s search of PubMed. Results: We will present the results of the searches, and the deduplication and screening of the records identified. If any additional trials are identified, we will inform the EMBASE project team. Conclusions: We will seek to draw conclusions on whether or not it is efficient for the Cochrane ENT TSC to continue to search EMBASE separately, how these data might inform the wider TSC community and future centralised searching projects and how any anticipated efficiency savings might impact on making skilled resources available for other Cochrane projects. Results: We will consider whether additional searching time and screening burden from supplementary PubMed searches generates unique included references. Conclusions: Ensuring Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Cochrane Database Syst Rev Suppl 1–327 (2015) 147 RO 23.9 Search summary table for systematic reviews (SRs) Bethel A1 , Rogers M1 University of Exeter, United Kingdom 1 Background: It is good practice in systematic reviews (SRs) to publish the search strategy used for each database, as recommended by the Cochrane Handbook for Systematic Reviews of Interventions. These can be used to inform future related reviews and update searches. However, it is currently not considered necessary to publish information about which databases held the relevant records and whether these were found by the search. Tracking and publishing this information as an additional search summary will aid future reviewers in choosing and using resources more effectively and will improve the efficiency of update searches. Objectives: To develop a search summary table that can easily be used at the end of all SRs to discover: 1. which databases the included references came from; 2. which other databases searched found the reference but it was excluded as a duplicate; 3. which other databases searched contained the reference but the search did not pick it up. Methods: The search summary was designed and developed by the authors over six systematic reviews including one Cochrane Systematic Review. These were then used to establish whether a future search on the same topic (or an updated search for the same review) could be carried out more efficiently with targeted database searching. Results: The results for each systematic review show, via a summary table, in which databases the included studies were found. An overall picture of the six summaries has also been produced to discover whether any pattern has emerged from the individual results. Conclusions: The search summary will inform future systematic reviews including updates to SRs or the update searches as part of the SR. This will provide evidence to the review team about which databases to search, particularly if there are time constraints. It will also be another step in sharing and refining search data. Background: Context and implementation of health interventions have received increasing attention over the past decade. Of particular interest is their influence on the effectiveness and reach of complex interventions. Yet, much confusion exists around these terms and their underlying concepts. Objectives: To analyze both concepts in a state-of-the-art assessment for the health sciences in order to create a common understanding of the use of ’context’ and ’implementation’ within systematic reviews and Health Technology Assessments (HTA). Methods: We performed two separate systematic searches for context (EMBASE,MEDLINE) and implementation (Google Scholar) to identify relevant theories, models and frameworks; 17 publications on context and 35 on implementation met our inclusion criteria. Pragmatic utility (PU) concept analysis was employed to analyze both concepts. PU comprises three guiding principles: selection of literature, organization and structuring of literature, and asking analytic questions of literature. Both concepts were analyzed according to four features of conceptual maturity, i.e. consensual definitions, clear characteristics, fully described preconditions and outcomes, and delineated boundaries. Results: Context and implementation were highly intertwined, influencing and interacting with each other. Context is defined as a set of characteristics and circumstances that surround the implementation effort. Implementation is conceptualized as a planned and deliberately initiated effort with the intention to put an intervention into practice. The concept of implementation presents largely consensual definitions and relatively well-defined boundaries, while distinguishing features, preconditions and outcomes are not fully articulated. In contrast, definitions of context vary widely, and boundaries with neighboring concepts, such as setting and environment, are blurred; characteristics, preconditions and outcomes are ill-defined. Conclusions: The maturity of both concepts should be improved further to facilitate operationalization in systematic reviews and HTA, e.g. by developing a framework that will allow an assessment of both concepts. RO 24.2 Rapid oral session 24 Novel approaches of interpreting and implementing healthcare evidence RO 24.1 Implementation and context: a concept analysis Pfadenhauer LM1 , Booth A2 , Burns J1 , Gerhardus A3 , Hofmann B4 , Lysdahl KB4 , Mozygemba K3 , Tummers M5 , Rehfuess EA1 1 Institute for Medical Informatics, Biometry and Epidemiology, University of Munich, Germany; 2 University of Sheffield, United Kingdom; 3 Institute of Public Health and Nursing Research, University of Bremen, Germany; 4 Centre for Medical Ethics, University of Oslo, Norway; 5 Radboud University Medical Center, The Netherlands Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Personalized music for residents with dementia in a long term care setting. Putting wheels on the evidence from the ALOIS database Hofstetter C1 1 Dementia and Cognitive Improvement Group, Canada Background: At Extendicare Guildwood, a community long-term care home in Toronto, the program director decided to expand the music program and employ the evidence that Dan Cohen utilized in his Music and Memory Project. Another dimension was the addition of the 4Life Foundation DJ Program youth to provide music expertise. Objectives: To provide personalized music for the residents on a regular basis and observe their reactions. To engage Cochrane Database Syst Rev Suppl 1–327 (2015) 148 the broader community, in particular ’at-risk’ youth, to create an intergenerational communication. Methods: A search of the ALOIS database for studies on music and dementia produced 56 studies offering evidence on the benefit of music therapy. A resident’s family donated iPods and CDs to Guildwood, which allowed a wide variety of music to be downloaded for the project. Personalized music folders were created for each resident by the 4Life Foundation volunteers. Results: Once weekly in group settings, residents experienced personalized music through headphones. The young men from the 4Life Foundation became part of the volunteer group at Guildwood and formed relationships with the music program participants. The residents who experienced the music intervention showed changes in affect/mood and changes in their demeanor and behaviour were observed by staff. Agitation in dementia patients was reduced during the intervention. Within 60 days of implementation of this project at the Guildwood site, Extendicare Canada committed to training and certifying the program directors in all 70 of their long-term care homes in Ontario to deliver the Music and Memory Project to their 10,000 residents. Conclusions: The studies in the ALOIS database are a valuable resource for improving care and quality of life in a long-term care setting. RO 24.3 The use of statistics as a drunken man uses lamp-posts: for support rather than illumination Corradini A1 1 Designskolen Kolding, Denmark Statistics is the science that studies the collection, analysis, interpretation and presentation of data. It makes it possible to learn, to extract patterns from data as well as to measure, and communicate uncertainty. The application of statistical methods to scientific, industrial, or societal problems allows trends to be determined in a specific context while also predicting what may happen in the future. It is therefore not surprising that in our fact-minded western cultures, the language of statistics is something very powerful to wield. When used correctly, statistical methods are very useful tools for analyzing what is happening in the world around us. On the other hand, statistics can easily be used to oversimplify a phenomenon or situation, thus ultimately resulting in confusing information, data misinterpretation, and deceived readers. Even worse, statistics can be purposely abused and misused in the attempt to sensationalize or inflate data reports and opinion polls related to social, economic, political, medical trends and conditions. Nonsense is what results when the data analysis is conveyed with dishonesty, with the wrong wordings or with the application of an incorrect statistical approach. Nonsense is also what results when the data analysis is presented to an audience that does not know what the statistics mean. Statistics plays an important role in the medical field and notably in the area of shared decision making. Many patients wish to know their individual risk of Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. being diagnosed with a certain condition or what are their chances of survival when they undergo a certain treatments. Explaining data to patients is however not an easy task. In this presentation, we review some use cases, which show how statistical data can easily be misinterpreted and, as such, should be avoided when using statistics for medical purposes. RO 24.4 Change score or follow-up score? An empirical evaluation of the impact of choice of mean difference estimates Fu R1 , Holmer H2 1 Pacific Northwest Evidence Based Practice Center, Department of Public Health and Preventive Medicine, Oregon Health, Science Univeristy, USA; 2 Department of Public Health and Preventive Medicine, Oregon Health, Science Univeristy, USA Background: In randomized controlled clinical trials, continuous outcomes are typically measured at both baseline and follow-up, and mean difference could be estimated by using the change score from baseline, using the follow-up scores, or using the analysis of covariance (ANCOVA) model. When there is baseline imbalance, the ANCOVA estimate is least biased, but often not reported. The impact of using the change versus the follow-up score has not been well studied. Objectives: Funded by the Agency for Healthcare Research and Quality, this study was to empirically assess the impact of using the change score versus the follow-up score on the conclusion of meta-analysis (MA). Methods: We included a total of 63 MAs (156 trials) from six comparative effectiveness reviews. We evaluated differences in baseline scores on the MA level and compared combined mean differences using the change score or the follow-up score. Discrepancy in conclusion occurs when one estimate (e.g. change score) shows significant difference and the other (e.g. follow-up score) does not. We also evaluated whether the impact varied qualitatively by alternative random effect estimates. Results: Based on the Dersimonian-Laird (DL) method, using the change score versus follow-up score led to five out of the 63 MAs (7.9%) showing discrepancy in conclusions; and based on the profile likelihood (PL) method, nine (14.3%) showed discrepancy. Using the change score was more likely to show a significant difference in effects between interventions (four out of five using the DL method, and seven out of nine using the PL method). The impact of the choice of the scores when using the restricted maximum likelihood method was similar to using the PL method. Using the Knapp-Hartung method led to most (10) MAs showing discrepancy. A significant difference in baseline scores did not necessarily lead to discrepancy in conclusion. Conclusions: Using the change score versus the follow-up score could lead to important discrepancies in conclusions. Sensitivity analyses using both change scores and final values should be conducted to check the robustness of results to the choice of mean difference estimates. Cochrane Database Syst Rev Suppl 1–327 (2015) 149 RO 24.5 Is Kudos a valuable resource for Cochrane authors? Stewart G1 , Hilton J2 , Woodward G1 1 Wiley, United Kingdom; 2 Cochrane Editorial Unit, United Kingdom Background: Kudos is an online platform that aims to help researchers explain, enrich and share their publications for greater research impact. Registered users of Kudos are able to ‘claim’ and create profiles for their published articles and increase readership and citations by adding links to resources, creating lay summaries and sharing information via social networks. In 2014, the publisher of the Cochrane Library, Wiley, took part in a trial of the Kudos platform. Wiley sent the details of over 25,000 published journal articles to Kudos. The details of all new Cochrane Reviews published in 2014 were also supplied. The contact author of every article was invited by Kudos to register and ‘claim’ their published work on the Kudos platform. The trial was extended to include all new and updated Cochrane Reviews published in 2015. Objectives: To establish how many Cochrane authors responded to the invitation from Kudos and what level of interaction there was with the Kudos platform. Why did some authors choose to use the Kudos platform and why did others ignore the invitation to use the platform? Did using the Kudos platform prove to be a positive experience for Cochrane authors and did they see increased levels of engagement in their work? Methods: Using data provided by Kudos we will establish the number of author registrations, the number of ‘claimed’ articles and the number of articles that have been enriched by Cochrane authors. This dataset will be compared to data from Wiley journal articles. A small survey of all Cochrane authors contacted by Kudos will be conducted. Two case studies will be used to provide an author’s perspective of using the Kudos platform. Results: In 2014, 1986 Wiley authors registered and 2233 articles were claimed by the registered authors. Initial results show that interaction among Cochrane authors was higher than among authors of other Wiley articles. Conclusions: Data to inform the conclusions will be gathered up until August 2015. This session will aim to establish whether the Kudos platform should be made more widely available to Cochrane authors as a platform to share the outcomes of Cochrane Reviews. RO 24.6 The use of content knowledge in the assessment of bias in trials Luijendijk H1 University Medical Center Groningen, The Netherlands 1 methods. Objectives: The aim of our paper is to show how reviewers can use their content knowledge to identify bias in trials. Methods: We introduce causal diagrams (DAGs) to illustrate how structural bias occurs as result of baseline differences, incorrect outcome measurement, and attrition. Selective reporting is, by exclusion, the fourth source of bias. Results: According to source of bias, we illustrate with examples how content knowledge allows the identification of (potential) bias, even if important trial methods such as randomization have been applied correctly. We also show how causal diagrams can help to understand the goals and proper use of trials methods. Conclusions: Reviewers can use their medical or other content knowledge to assess sources of bias in trials. Causal diagrams complement the definitions for correct application of trial methods in current assessment tools. RO 24.7 Cochrane Reviews in dentistry: making sense of evidence at its highest level –an analysis Mulimani P1 1 Melaka Manipal Medical College, Malaysia Cochrane reviews are considered to provide scientific evidence at its highest level by processing all available data regarding a research question through rigorous tests and standardized exacting specifications. Often the methodology, to those not involved or to those viewing the information from outside, appears complex or technical. Cochrane itself has made provisions including plain language summaries, Cochrane Clinical Answers, participation of consumers in Review Groups or forums, to make the information access easier and more meaningful to stakeholders including lay-people, patients, care-givers and clinicians. This paper provides an analysis of the various channels that are, and could be, employed to convey and disseminate the importance of high-quality evidence and to interpret and translate the findings of Cochrane research into understandable, clinically relevant and practically applicable information of utility to patients and doctors in the field of dentistry. Additionally this paper will also discuss how the importance of evidence, awareness of levels and quality of evidence and significance of Cochrane Reviews and methods can be made more pervasive and wide-spread among the healthcare professionals through continued professional development programs. We explore exposure to these during the formative years of their professional lives, as students, through various teaching-learning activities, thus creating a newer generation of health professionals who are more research and evidence-savvy. Background: Current assessment tools focus on the methods that are commonly used in trials to reduce bias. However, these methods minimize but do not guarantee lack of bias. It is important that sources of bias are assessed too. Moreover, many reviewers do not have in-depth knowledge of trial Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Cochrane Database Syst Rev Suppl 1–327 (2015) 150 RO 24.8 A systematic review and external validation of prediction models for gestational diabetes: the RESPECT study de Ruiter M1 , Kwee A1 , Naaktgeboren CA2 , Franx A1 , Moons KGM2 , Koster MPH1 1 Department Women and Baby, UMC Utrecht, The Netherlands; 2 Julius Center for Health Sciences and Primary Care, UMC Utrecht, The Netherlands Background: The rising incidence of gestational diabetes (GDM) contributes to an increasing number of adverse pregnancy outcomes (for example macrosomia and neonatal hypoglycemia). Currently, risk stratification in early pregnancy is based on single medical and/or obstetrical history risk factors. GDM, commonly diagnosed in late pregnancy, is particularly suited for early prediction in pregnancy. Timely recognition and treatment of GDM will offer opportunities to improve pregnancy outcome. Objectives: Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. To identify and validate existing prediction models for GDM in the first trimester of pregnancy. Methods: MEDLINE and EMBASE were searched up to December 2014. First trimester non-invasive prediction models for GDM in current pregnancy were included as well as external validation studies thereof. Eligibility was appraised by two reviewers. For each prediction model, data were extracted according to the CHecklist for critical Appraisal and data extraction for systematic Reviews of prediction Modelling Studies (CHARMS). Models were validated using data from a large prospective study, the RESPECT cohort (3641 pregnancies of which 171 developed GDM). Results: Our systematic review identified 20 articles, 16 of which were model development studies and four were external validations. Of the 16 models, 12 could be validated (Figure 1). In developmental studies C-statistics ranged from 0.63 to 0.82. Four models have been externally validated with C-statistics from 0.60 to 0.76. In our validation study, 11 recalibrated models yielded C-statistics of 0.67 to 0.78. At a fixed false positive rate of 10%, sensitivity ranged from 27% to 43%. Most of the models showed acceptable to good calibration (Figure 2). Common predictor variables included Cochrane Database Syst Rev Suppl 1–327 (2015) 151 Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Cochrane Database Syst Rev Suppl 1–327 (2015) 152 in the models were age, body mass index, ethnicity, history of GDM and family history of diabetes mellitus. Conclusions: Prediction models for GDM can be used to identify high risk pregnancies. This allows early risk stratification and enables customized targeted interventions. Screening will contribute to the development of personalized obstetric care and improve the utilization of health care resources. The results of sensitivity analyses showed only three (0.1%) reviews’ pooled effect sizes were affected by the data of those studies. Conclusions: Most Cochrane Reviews have searched the trial register, and small number of reviews include the studies from registers. However, few reviews’ results are affected by the studies from registers. LRO 9.2 Lunch rapid oral session 9 Trial registration LRO 9.1 The impact of studies from a trial register for the results of systematic review: a survey of Cochrane reviews Wei D1 , Qu Q2 , Chang X3 , Song X4 , Gao Y4 , Zhang Y4 , Ling N4 , Chen Y1 , Yang K1 1 School of Basic Medical Sciences, Lanzhou University; Key Laboratory of Evidence-Based Medicine and Knowledge Translation of Gansu Province, Lanzhou, China; 2 The Second Clinical Medical College of Lanzhou University, Lanzhou, China; 3 The First Clinical Medical College of Lanzhou University, Lanzhou, China; 4 School of Basic Medical Sciences, Lanzhou University, Lanzhou, China Background: There are rigorous requirements for each step of a systematic review in the Cochrane Handbook for Systematic Reviews of Interventions. For instance, there must be two independent reviewers when screening eligible studies and extracting the data, which supports robust evidence. Furthermore, it suggests that the trial register should be searched for each Cochrane Review. To date, there are more 20 international trial registers, such as the World Health Organization International Clinical Trials Registry Platform (WHO ICTRP), ClinicalTrials.gov, etc. However, is it really significant to search the trial register for systematic review? Objectives: To investigate the search of trial register in Cochrane Reviews and analyze whether it would change the results of the reviews. Methods: We sampled the Cochrane Reviews published in 2013. Two reviewers extracted the data independently. The extracted data included whether it had searched trial register, the number and name of trial register, the trials included in reviews, and whether the researchers could achieve the data, etc. We assessed the impact of studies from trial register for the pooled effect size of the reviews through sensitivity analyses. Results: There were 992 Cochrane reviews published in 2013. Of those, 974 (98.2%) had searched the registers (mean = 2, range: 1 to 20). The top five frequently searched registers were Cochrane Group Registers (91%, 890/974), ClinicalTrials.gov (43%, 423/974), WHO ICTRP (34%, 331/974), CCT (19%, 181/974), and the metaRegister of Controlled Trials (mRCT; 12%, 120/974). Thirty-two (3%) reviews included the studies from registers, and nine (1%) reviews synthesized the data from studies from registers. Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Comparison of the attention paid to trial registration in the Asia-Pacific region in the last three years Tam KW1 , Huang TW2 , Kuan YC1 1 Taipei Medical University, Shuang Ho Hospital, Taiwan; 2 HungKuang University, Taiwan Background: Trial registration is known to improve research transparency and will ultimately strengthen the validity and value of the scientific evidence base. More and more biomedical journals support clinical trial registration policies after the requirement for registration was introduced by several medical journals, led by the International Committee of Medical Journal Editors (ICMJE). However, the attention paid to trial registration varies among countries, particularly in the Asia-Pacific Region. Objectives: To evaluate the attention paid to trial registration in selected countries in the Asia-Pacific Region in the last three years. Methods: The published randomized controlled trials (RCTs) from each country were extracted by restricting the ’affiliation’ field in PubMed searches to individual country names. The selected countries or regions consisted of Japan, South Korea, Singapore, Hong Kong and Taiwan. Past (October 2011 to March 2012) and present (June 2014 to January 2015) periods were searched. We then compared the performance of each country with regard to RCTs in terms of trial registration. Results: We identified that from the five countries and regions included in the study, 29.4% (past) and 85.2 % (present) of the RCTs produced by Hong Kong had trials that were registered, followed by Singapore (18.2% in the past; 77.8% at present), Taiwan (13.9% in the past; 58.6% at present), Japan (26.9% in the past; 53.6% at present), and South Korea (22.8% in the past; 51.1% at present). Conclusions: In the Asia-Pacific Region, trial registration as part of the current research paradigm is of greater concern in the present period. However, the proportion of trials that are registered is still not satisfactory. Our results suggest that researchers should pay more attention to registering their studies on the registry platform. Cochrane Database Syst Rev Suppl 1–327 (2015) 153 LRO 9.3 LRO 9.4 Tuberculosis clinical trial activity on the African continent: an analysis of tuberculosis trials registered on the Pan African Clinical Trials Registry and WHO-ICTRP How complete are prostate cancer studies conducted in Germany as displayed in the national trial register of the Deutsche Gesellschaft für Urologie (German Association of Urology)? Pienaar E1 , Kredo T1 , Abrams A1 , Lutje V2 1 SACC, South Africa; 2 Cochrane Infectious Diseases Review Group, South Africa Background: Thirty-four per cent of tuberculosis (TB)-related deaths occur in Africa. Among people living with HIV, at least one in four deaths can be attributed to TB, often in resource-limited settings. The Pan African Clinical Trials Registry (PACTR), the only African member of the World Health Organization’s (WHO) Primary Registry Network, contributes to WHO’s International Clinical Trial Registry Platform’s (ICTRP) central database of clinical trials as well as initiatives to strengthen regulatory and research frameworks by providing a venue and support for trial registration. This study describes the status of TB trials registered on PACTR and the WHO-ICTRP. Objectives: To analyse TB interventions trials conducted in Africa and registered on PACTR and WHO-ICTRP. Methods: We searched PACTR and WHO-ICTRP for trials of TB interventions. Data extraction included the number of trials, their location, intervention studied, principal investigator’s (PI) location, participant age range and funder. Descriptive analysis was run in MS Excel. Results: There are 430 studies registered on PACTR, with 44 trials reporting planned or ongoing TB research. On ICTRP 173 intervention trials were found as well as 25 TB-related observational studies. Twenty-four PACTR trials explored treatment, seven diagnosis, five prevention trials included two vaccines, four education and counselling, and two supportive care interventions. Sample sizes range from 12 to 27 000. Twenty-one trials were single-centred in 10 countries; 23 multi-centre trials in 10 African countries; 38 PIs are African, one trial had multiple PIs - nine from Europe and USA. Nine trials received funding from the European and Developing Countries Clinical Trials Partnership (EDCTP); none were industry funded. Further analysis to include and compare with trials registered on WHO-ICTRP is ongoing. Conclusions: As PACTR registrations increase, the registry will become a tool for regulatory and funding bodies to monitor the changing TB trial landscape in Africa. Both PACTR and WHO-ICTRP are valuable resources for systematic review authors to identify ongoing TB trials, and also potential new review titles. Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Schmidt S1 , Becker C1 , Jena S2 , Wullich B3 , Kunath F3 1 DGU, Germany; 2 DRKS, Germany; 3 Department of Urology, University Hospital Erlangen, Germany Background: The German Association of Urology (DGU) initiated a project in 2010 to easily register clinical trials within the German Clinical Trials Register (DRKS), named DGU sub-register. The DRKS is a primary register of the World Health Organization (WHO) International Clinical Trials Registry Platform (ICTRP). Urology-related studies were retrospectively included in the DGU sub-register. However, the completeness of urologic trial registration in the DGU register remains unclear. Objectives: To assess whether prostate cancer studies conducted in Germany were registered in the DGU sub-register. Methods: The WHO ICTRP database was searched for all prostate cancer trials conducted in Germany. The retrieved registration identifiers were used as search terms to identify the associated trials registered in the DGU sub-register. The increase in newly registered trials per year was analyzed using linear regression. Results are presented with 95% confidence intervals. A P value less than 0.05 was considered statistically significant. Results: The search in WHO ICTRP identified 872 records for 194 trials. The annual number of registered trials identified in WHO ICTRP did not increase significantly (P value 0.063). Of all the identified trials, 108 (56%) also had a DGU-DRKS identifier and 25 (13%) were primarily registered in the DGU sub-register. While 67% of the trials were registered in 2010 to 2015, 33% were registered in 2009 or earlier. The annual increase was statistically significant (0.89 more registered trials/year; 95% CI 0.18 to 1.59: P value 0.018). Conclusions: Only half of the prostate cancer studies are displayed in the DGU sub-register. Since the establishment of the DGU sub-register, trial registration increased annually. Whether this development is due to the availability of a user-friendly registration platform or to an increased awareness of trial registration cannot be determined. Furthermore, determining the true proportion of registered studies remains a challenge. Results may be biased by the completeness and quality of entries in the WHO ICTRP. Keyword: ‘prostate cancer’, limitation: ‘recruitment location’ Cochrane Database Syst Rev Suppl 1–327 (2015) 154 LRO 9.5 LRO 9.6 Prematurely discontinued randomized trials are frequently labelled ‘completed’ in trial registries: a systematic review Epidemiology and reporting characteristics of systematic review protocols Schandelmaier S1 , Al Turki R1 , Olu KK1 , Frei R1 , Von Niederhäusern B1 , Agarwal A2 , Briel M1 1 University Hospital, Basel, Switzerland; 2 McMaster University, Hamilton, Canada Background: Trial registries such as clinicaltrials.gov typically include information about the status of a randomized controlled trial (RCT), i.e. whether it is ongoing, completed, or discontinued. Trial registries can be a useful resource for systematic reviewers who are searching for unpublished RCTs, such as ongoing or often discontinued RCTs. The accuracy of trial status information in registries has never been investigated. Objective: To examine whether trial registries accurately label discontinued RCTs as ‘discontinued’ and provide reasons for RCT discontinuation; and to identify factors associated with inaccurate trial status information. Methods: Systematic reviews of published RCTs reported as discontinued and registered. We systematically searched MEDLINE and EMBASE (from 2010 to 2014). Pairs of reviewers independently extracted data from publications and corresponding registries using pre-piloted data extraction forms. We performed multivariable regression to identify risk factors for inaccurate trial status information. Results: We included 173 discontinued RCTs that were registered in various trial registries, most frequently in clinicaltrials.gov (77%). RCTs were mostly discontinued due to slow recruitment (62%), futility (19%), or harm (17%). Of the 173 discontinued RCTs, 45% were accurately labelled as discontinued; the remaining RCTs were wrongly labelled as completed (40%), suspended (9%), on-going (5%), or not started yet (1%). Most RCTs that were accurately labelled also provided the accurate reason for discontinuation (57/77, 74%; Table 1). Significantly, accurate registration of trial status became more frequent over time (adjusted odds ratio 1.16 per year; 95% confidence interval 1.04 to 1.3) whereas trial features such as industry sponsorship, multiple centers, or sample size were not significantly associated with accurate trial status information. Conclusion: Less than half of published discontinued RCTs were accurately labelled as discontinued in registries. Systematic reviewers searching trial registries for eligible trials should view trial status information with caution. Shamseer L1 , Page M2 , Reid E3 , Turner L4 , Sampson M5 , Moher D1 1 Ottawa Hospital Research Institute; University of Ottawa, Canada; 2 School of Public Health and Preventive Medicine, Monash University, Australia; 3 Department of Pharmacy, Vancouver General Hospital, Canada; 4 Ottawa Hospital Research Institute, Canada; 5 Children’s Hospital of Eastern Ontario, Canada Background: Protocol development is a fundamental preliminary step in the systematic review process, enabling a priori consideration of review methods and preventing ad hoc decision-making during the review process. When available, protocols can facilitate detection of selective reporting when compared to completed reviews. While protocols are a mandatory component of Cochrane Reviews, most reviews are done outside Cochrane and few report working from protocols. The reporting completeness of systematic review protocols has not previously been examined. Objectives: This study will assess: the proportion of systematic reviews that have been registered and/or have an available protocol; and for those with a protocol, the epidemiological characteristics and completeness of reporting. Methods: Three-hundred systematic reviews indexed in MEDLINE in February 2014, 14% of which are Cochrane Feviews, form the sample from which protocols and registration information will be sought. We will search PROPSERO to determine the number of registered reviews (post-October 2013 only for Cochrane Reviews) and will use a systematic search to identify whether a public protocol exists. For all identified protocols, epidemiological characteristics (e.g. number and country of authors, ICD-10 category) will be extracted and reporting characteristics will be assessed based on items from the PRISMA extension for protocols. Data will be extracted by a single assessor after meeting a minimum threshold of agreement (80%) during independent pilot-testing by two assessors. We will summarise data as proportions for dichotomous items and median and interquartile range for continuous items. We will compare reporting between Cochrane versus non-Cochrane protocols and registered versus non-registered protocols. Results: Results will be available for presentation at the Colloquium. Significance: This study will provide data on the prevalence of review registration, availability of review protocols, and the completeness of information reported in SR protocols. This study will be the first of its kind to provide evidence on the strengths and weaknesses of SR protocol reporting. Table 1 Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Cochrane Database Syst Rev Suppl 1–327 (2015) 155 Lunch rapid oral session 10 Missing data LRO 10.1 How much study participant data is missing from our trials? Kirkham J1 , Dwan K2 , Blümle A3 , von Elm E4 , Williamson P1 1 University of Liverpool, United Kingdom; 2 Cochrane Editorial Unit, United Kingdom; 3 German Cochrane Center, Germany; 4 Cochrane Switzerland, Switzerland Background: Study publication bias and outcome reporting bias (ORB) have been recognised as two threats that can affect the validity of systematic reviews due to problems with missing outcome data. There is strong evidence of an association between significant results and publication; studies that report positive or significant results are more likely to be published and outcomes that are statistically significant have higher odds of being fully reported [1]. Objectives: To estimate the proportion of missing data due to lack of publication of the studies and the proportion of missing outcome data within a published study. Methods: Data were taken from protocols of randomised controlled trials submitted to the research ethics committee of the University of Freiburg (Germany) between 2000 and 2002 and associated full published articles. The total amount of published and unpublished outcome data from all study participants was extracted and the proportion of missing data from both unpublished and published studies computed. Missing data from unpublished studies was taken from planned sample sizes from study protocols. A sensitivity analysis was undertaken to account for partially reported outcome data. Results: From 259 studies in the study cohort, 167 were published and 92 were unpublished. Half (51%; 1,288,719 of 2,522,010) of the participant outcome data from the 259 studies was missing; 39% of this missingness was attributable to missing data from published studies and 12% from unpublished studies. The sensitivity analysis revealed that the majority of missing data from published studies was linked to data being reported in a way that could not be included in a meta-analysis. Conclusions: Missing participant outcomes data from both published and unpublished studies is frequent. Preventative measures and potential statistical solutions for reducing the impact of missing data in meta-analyses will be discussed. Reference [1] Dwan K M, Gamble C, Williamson P R, Kirkham J J. Systematic review of the empirical evidence of study publication bias and outcome reporting bias - an updated review. PLoS ONE 2013: 8: e66844. Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. LRO 10.2 Extent and handling of missing dichotomous outcome data for participants in 672 trials included in 100 Cochrane and non-Cochrane systematic reviews Kahale LA1 , Diab B1 , Mustafa R2 , Busse JW3 , Agarwal A3 , Waziry R1 , Li L4 , Lopes LC5 , Koujanian S6 , Dakik A1 , Chang Y3 , Guyatt GH3 , Akl EA1 1 American University of Beirut, Lebanon; 2 University of Missouri-Kansas City, USA; 3 McMaster University, Canada; 4 Sichuan University, China; 5 University of Sorocaba, Brazil; 6 Lebanese American University, Lebanon Background: Reports of randomized clinical trials (RCTs) do not always specify whether participants belonging to certain categories (e.g. those who withdraw consent, non-compliers) were followed-up for the outcome(s) of interest or not (i.e. missing participant data (MPD)). Systematic reviewers may, therefore, have to make assumptions on whether data from these participants is missing. Objectives: To describe characteristics, reporting and handling of categories of participants that could be potentially counted as having MPD in RCTs included in Cochrane and non-Cochrane systematic reviews (SRs). Methods: We selected a random sample of 100 Cochrane and non-Cochrane SRs published in 2012. Eligible SRs reported a group-level meta-analysis of a patient-important dichotomous efficacy outcome, with a statistically significant effect estimate. Ten reviewers, working in pairs, independently extracted data from the RCTs included in eligible meta-analyses. We focused on 10 categories that could be potentially counted as MPD: ’participants deemed ineligible after randomization’, ’did not receive any treatment’, ’withdrew consent’, ’outcome not assessable’, ’explained and unexplained lost to follow-up’, ’dead’, ’experienced adverse events’, ’non-compliant’, ’discontinued prematurely’, and ’cross-over’. We specifically assessed: (1) the number of participants in those categories; (2) whether trialists explicitly reported on the follow-up status of those participants; and (3) how trialists handled those participants in their analysis. Results: We included a total of 672 RCTs included in 100 eligible Cochrane and non-Cochrane SRs. We are in the process of data abstraction. We will present the findings of the study at the Colloquium. Conclusions: Results will inform a better understanding of trial reporting on the follow-up, and handling of categories with potentially missing outcome data, an issue of importance for systematic reviewers and for standards of trial reporting. Cochrane Database Syst Rev Suppl 1–327 (2015) 156 LRO 10.3 LRO 10.4 Reporting and analysis of missing participant continuous data in randomized controlled trials (LOST-IT-II): a systematic survey Reported versus actual analytical methods to handle missing participant data in meta-analysis: a survey of 100 systematic reviews Zhang Y1 , Alexander Kennedy S2 , Abu Bakar Aloweni F3 , Flórez ID4 , Colunga L5 , Craig S6 , Zhang S7 , Li A1 , Cruz Lopes L8 , Fei Y9 , Alexander P1 , J. Riva J10 , Sadeghirad B1 , Agarwal A11 , Akl E12 , Schünemann HJ13 , Thabane L1 , Guyatt G13 1 Department of Clinical Epidemiology and Biostatistics, McMaster University, Canada; 2 Department of Medicine, McMaster University, Canada; 3 Singapore General Hospital, Singapore; 4 Universidad de Antioquia, Medellin and McMaster University, Colombia, Canada; 5 Hospital Civil de Guadalajara, ”Fray Antonio Alcalde”, México; 6 Michael G DeGroote Institute for Pain Research and Care, McMaster University, Canada; 7 Medical Affairs, GSK, Canada; 8 Universidade de Sorocaba, São Paulo, Brazil; 9 Beijing University of Chinese Medicine, McMaster University, China, Canada; 10 Department of Family Medicine, McMaster University, Canada; 11 Faculty of Medicine, University of Toronto, Canada; 12 Clinical Epidemiology Unit and Center for Systematic Reviews in Health PolicyandSystemsResearch(SPARK),AmericanUniversityofBeirut, Lebanon; 13 Department of Medicine and Clinical Epidemiology and Biostatistics, McMaster University, Canada Background: Missing participant data (MPD) can bias trial results and conclusions if the missingness is associated with the occurrence of the outcome of interest. No study has summarized the reporting and analytic approaches of MPD that authors have used for continuous outcomes in randomized controlled trials (RCTs). Objective: Our objectives are to assess: (1) how authors report MPD for continuous outcomes; (2) the analytic methods used for primary analysis and sensitivity analysis to address MPD; and (3) the impact of these sensitivity analyses on the conclusions. Method: We are conducting a systematic survey of RCTs published in 2014 in core medical journals. We include RCTs reporting at least one patient-important outcome analyzed as a continuous variable. We randomly sample RCTs aiming at a sample size of 200. We will calculate the proportion of RCTs that explicitly reported whether MPD occurred, percentage of randomized participants with MPD for the continuous outcome, the methods used to handle MPD in the primary and sensitivity analyses, and how often the conclusions are interpreted considering the sensitivity analyses conducted and whether consideration of MPD changes the conclusion. We will conduct regression analyses with independent variables general trial characteristics (e.g. type of funding, type of intervention) and methodological trial characteristics (e.g. allocation concealment, length of follow-up), to establish factors associated with reporting, analysis, and conclusions. Conclusion: Our methodological survey will have important implications for both trialists and users of trial evidence in interpretation of the findings from RCTs with MPD on continuous outcomes. Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Kahale L1 , Diab B1 , Mustafa R2 , Busse JW3 , Agarwal A3 , Waziry R1 , Li L4 , Lopes LC5 , Koujanian S6 , Dakik A1 , Chang Y3 , Guyatt GH3 , Akl EA1 1 American University of Beirut, Lebanon; 2 University of Missouri-Kansas City, USA; 3 McMaster University, Canada; 4 Sichuan University, China; 5 University of Sorocaba, Brazil; 6 Lebanese American University, Lebanon Background: Systematic review authors do not typically report the analytical methods they used to handle missing participant data (MPD) in meta-analyses. Moreover, when methods are reported, the concordance with reviewers’ actual analyses is uncertain. Objectives: To determine which analytical methods systematic review authors actually used to deal with MPD in their meta-analyses of dichotomous outcomes, and whether there is concordance with the approach described in their methods section. Methods: Eligible systematic reviews (SRs) reported a group-level meta-analysis of a patient-important dichotomous efficacy outcome, with a statistically significant effect estimate. We acquired the full texts of all trials included in the meta-analyses of interest. Ten reviewers working in pairs will independently abstract statistical data relevant to the meta-analysis of interest from both the SR (numerator, denominator) and the corresponding RCTs (number randomized, number with missing data, number of observed events). Then, we will compare the SR data to the corresponding trial data. Based on these comparisons, we will classify the analytical method the SR authors actually used to deal with MPD into four categories: ’complete case analysis’, ’making assumptions for MPD’, ’using the trialists’ assumptions’, or ’excluding trials with high rate of MPD’. Then, we will assess the consistency of the method across trials included in the same meta-analysis, and concordance with the method SR authors report to have used. Results: We included a random sample of 100 Cochrane and non-Cochrane SRs published in 2012. We identified a total of 672 RCTs included in these SRs. We are in the process of data abstraction. We will present the findings of the study at the Colloquium. Conclusions: The findings of our study are likely to reinforce the importance of transparency in reporting and conducting SRs, particularly when dealing with MPD. Cochrane Database Syst Rev Suppl 1–327 (2015) 157 LRO 10.5 Prognostic models for cardiovascular disease risk in the general population: a systematic review Damen J1 , Hooft L1 , Schuit E1 , Debray T1 , Collins GS2 , Tzoulaki I3 , Lassale CM4 , Siontis G5 , Black J6 , Heus P1 , van der Schouw YT7 , Peelen L7 , Moons K1 1 Dutch Cochrane Centre, Julius Center for Health Sciences and Primary Care, University Medical Center Utrecht, The Netherlands; 2 Centre for Statistics in Medicine, NDORMS, Botnar Research Centre, University of Oxford, United Kingdom; 3 Department of Epidemiology and Biostatistics, Imperial College of Medicine, London, United Kingdom; 4 Department of Epidemiology and Biostatistics, School of Public Health, Imperial College London, United Kingdom; 5 Department of Cardiology, Bern University Hospital, Switzerland; 6 MRC Epidemiology Unit, University of Cambridge School of Clinical Medicine, United Kingdom; 7 Julius Center for Health Sciences and Primary Care, University Medical Center Utrecht, The Netherlands Background: Cardiovascular disease (CVD) is a leading cause of morbidity and the leading cause of mortality worldwide. Many prediction models have been developed to assess individual CVD risk to allow targeting of preventive treatment. Objectives: To give an overview of all prognostic models that predict future risk of CVD in the general population, and to describe predicted outcomes, study populations and included predictors. Methods: In June 2013 a systematic search was performed in MEDLINE and Embase to identify studies that described the development, validation or incremental value of a multivariable prognostic model predicting CVD in the general population. Results: Out of 9671 papers identified, 314 were included, describing the development of 373 prognostic models, 519 external validations and 278 incremental value assessments. In total 132 models (35%) were externally validated, of which 70 models (19%) were validated by independent researchers. Most models were developed in Europe (n = 258), and predicted risk of coronary heart disease (n = 115, Table 1) over a 10-year period (n = 211). Furthermore, most prevalent predictors were age and smoking (n = 323 and n = 332 respectively, Figure 1), frequently with separate models for males or females (n = 256). Discrimination and calibration were reported for only 62% and 52% of the external validations respectively. Crucial clinical and methodological information was often missing, and when reported, substantial heterogeneity in predictor and outcome definitions was seen between models. For 53 models the prediction time horizon was not reported and for 56 models the intercepts or baseline hazards were not reported, making it impossible to use for individual risk prediction. Conclusion: In a similar manner to various other clinical domains, there is an excess of prognostic models predicting CVD in the general population. Additionally, the usefulness of most models remains unclear, due to incomplete presentation, lack of external validation, and heterogeneity in predicted outcomes and study populations. Future research should focus on validating, updating, meta-analyzing and studying impact of models. Table 1 Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Cochrane Database Syst Rev Suppl 1–327 (2015) 158 Figure 1 LRO 10.6 Meta-analysis methods for better validation of prediction models Steyerberg E1 1 Erasmus MC, The Netherlands Background: Prediction models are increasingly developed for many diagnostic and prognostic endpoints, and increasingly based on individual patient data from multiple studies. Validation of predictions from such models is important, but many researchers use inefficient designs and analyses. Objectives: We aimed to evaluate the role of meta-analytic methods in the validation of prediction models. Methods: We considered 15 studies of patients with traumatic brain injury (n = 11,026), where we predicted 6-month mortality. Prediction models were constructed in each of the 15 studies, and in the pooled data set. Various approaches to internal and external validation were explored. Results: Logistic regression models included 10 predictors, with good overall discriminatory ability (median apparent c Figure 1 Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Cochrane Database Syst Rev Suppl 1–327 (2015) 159 statistic in the 15 studies 0.81). For internal validation per study, we confirmed that a random split-sample approach was inefficient and unstable, in contrast to bootstrap resampling. In the pooled data set, an internal-external validation procedure provided insight in the substantial variability in discriminatory ability (c statistic) by study, as well as variability in calibration (study-specific intercept and calibration slope). This variability could well be quantified by random-effects meta-analysis modeling and I2 (I-squared) statistics (> 50%). The predictive distribution of predictions for patients in future studies could hence be estimated, and was much wider than that based on a fixed-effect analysis in the pooled data set. Conclusions: We conclude that meta-analytic methods are very useful to obtain insight in the validity of prediction models. Quantification of heterogeneity in predicted risks should become the focus of validation studies, rather than confirmation of performance. tests and comparisons by each study) and quality assessment, as well as to generate paper-based data extraction- and reporting-forms. (See Fig. 1) Results: We present the feasibility of a customized interactive database using the example of an ongoing Cochrane DTA review. Conclusions: Based on tools that are already available, a custom-made interactive database to fulfill the special needs of DTA-collaborators proved successful and productive. This tool will be made available on the ACE website to be used by other groups with similar needs. LRO 11.2 The reference standard for diagnostic imaging test accuracy Reed M1 1 Canadian Association of Radiologists, Canada Lunch rapid oral session 11 Diagnostic test accuracy reviews LRO 11.1 Generating an interactive database for international collaboration on a Cochrane diagnostic test accuracy-review Roth D1 , Lee A2 , Hovhannisyan K3 , Hartmann A1 , Pace NL4 , Herkner H1 1 Medical University of Vienna, Department of Emergency Medicine, Austria; 2 Department of Anaesthesia and Intensive Care, The Chinese University of Hong Kong, Hong Kong; 3 Rigshospitalet, The Cochrane Anaesthesia Review Group, Copenhagen, Denmark; 4 Department of Anesthesiology, University of Utah, Salt Lake City, USA Background: Reviews of diagnostic test accuracy (DTA) are increasingly gaining importance within Cochrane, and Cochrane tools, such as RevMan, are constantly being updated to serve the requirements of conducting such types of review. However, not all tasks can currently be solved using standard Cochrane tools. Objectives: The authors of the DTA-review ’Airway physical examination tests for detection of difficult airway management in apparently normal patients’ of the Cochrane Anaesthesia, Critical and Emergency Care Group (ACE) strived for a tool to facilitate data-collection for their team scattered across four countries and three continents. The need for such a tool was aggravated by the fact, that this project represents a ’multidimensional’ review with multiple target conditions, a multitude of tests, serving both as standard- and index-tests, and that each individual study usually includes a large number of comparisons. Methods: Based on the QUADAS 2 database by Whiting et al, an MS Access-database to be hosted by a filehosting-service was developed to collect data on measures of DTA on multiple levels (multiple target conditions, index Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Background: The reference standard for measuring diagnostic test accuracy ideally should be error-free in establishing the diagnosis of the condition of interest and should be applicable to all of the population being tested. Unfortunately it is virtually impossible to find such reference standards to measure the accuracy of diagnostic imaging (DI) tests. Objective: the objective of this presentation is to describe and illustrate the principal barriers to finding an ideal reference standard for assessing DI test accuracy. Analysis: there are three main barriers to finding an ideal reference standard for assessing the accuracy of DI tests. 1. The high quality of current DI technology. Pathology is usually considered to be the ideal reference standard for assessing DI test accuracy. However, DI very often provides imaging of pathology which is as good as anatomical pathology. Common examples would be fractures, gallstones, appendicitis and intracranial hemorrhage. 2. The reference standard cannot ethically be applied to all study participants. One of the most common uses for DI is to diagnose conditions, such as tumours, injuries or abscesses, which require surgical intervention. The findings at surgery are an excellent reference standard. However, surgery can only ethically be performed on those patients with a positive DI test. 3. There is no practical reference standard for the condition being assessed. Examples of this would include such common conditions as pneumonia and pulmonary oedema for which DI is one of the primary diagnostic tests. Conclusion: finding a reference standard for assessing DI test accuracy is extremely difficult. Methodologists and radiologists need to work together to establish appropriate criteria for determining reference standards for assessing DI test accuracy. Cochrane Database Syst Rev Suppl 1–327 (2015) 160 LRO 11.3 Methods for assess validity of diagnostic tests in rare diseases Canon L1 , Perez A1 , Avellaneda P1 , Serrano P1 1 Instituto de Evaluación Tecnológica en Salud, Colombia Background: Rare diseases are those with a prevalence of 1 per 5000 individuals or, within the European Union, fewer than 5 per 10,000 persons. According to Colombian law there should be full access to diagnostic technologies for rare diseases in our country. In order to prioritize them, 11 reviews of diagnostic validity were developed, upon which this paper is based. Objectives: To describe the findings of methodology for the development of diagnostic validity reviews for rare diseases. Methods: Eleven assessments of diagnostic utility tests were done in Colombia in 2014 to perform this descriptive study. The most important aspects in the development of reviews for low-prevalence diseases were identified with little evidence available. Results: 72% of assessed diseases had specific prevalence data for Colombia. The built-in assessments evidence was 9% of systematic literature reviews, 27.2% of cohort studies, 18% of diagnostic test studies and 36.3% of descriptive studies. No evidence was identified to answer the research question in 27% of the 11 assessments done. The quality of the studies was assessed with AMSTAR (A MeaSurement Tool to Assess systematic Reviews; 9%) and QUADAS 2 (Quality Assessment of Diagnostic Accuracy Studies tool; 36%). Due to the study type 45% of the evidence did not have quality assessment. In two of the 11 assessments done combined effect analysis was performed. Conclusions: Despite the limited evidence for rare diseases, systematic and reproducible evaluations can be done in order to have an input for decision-making in the countries. The prevalence and pathophysiology regarding rare diseases do not allow the performance of studies of high methodological quality, however, the evidence from descriptive studies is quite useful in this context. Given the limitations of the sparse evidence from observational studies and descriptive designs it must be reported with high quality, which means a challenge for the production of systematic literature reviews. LRO 11.4 Summarising diagnostic accuracy using a single parameter Simmonds M1 for Reviews and Dissemination, University of York, United Kingdom 1 Centre Background: In meta-analyses of diagnostic test accuracy, results are often presented as separate analyses of sensitivity and specificity, but this ignores the correlation between these parameters. More complex approaches such as the bivariate approach or hierarchical summary receiver operating characteristic (HSROC) curve model do account for the correlation, but are more difficult to implement and interpret. Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Objectives: To demonstrate how diagnostic accuracy may be more simply summarised using a single parameter rather than sensitivity and specificity. Summarising the diagnostic accuracy of a test in a study with a single parameter permits the use of standard meta-analysis techniques for pooling results across studies, making interpretation simpler. Four parameters will be considered: diagnostic odds ratio, diagnostic standardised mean difference, Lehmann’s index and Youden’s index. Methods: The four parameters will be described, and shown to be related to the underlying distribution of the diagnostic test. It will be shown how all four parameters may easily be estimated using simple regression models. The four new parameters will be compared, and compared to existing methods, using commonly cited examples of diagnostic test accuracy meta-analyses. Results: As will be seen from various examples, the four single parameter methods offer a generally robust alternative to the HSROC model. Summary results and conclusions can vary substantially between methods, demonstrating that summaries of diagnostic accuracy may be sensitive to assumptions about the underlying distribution of the diagnostic test. Conclusions: Summarising diagnostic accuracy using one of the proposed single parameter methods offers an alternative approach to more complex methods such as the HSROC model. Results of these analyses can easily be presented on a forest plot, making interpretation of results simpler and making interpretation of heterogeneity and comparisons between different diagnostic tests straightforward. LRO 11.5 Using full information on cut-offs per study to determine an optimal cut-off in diagnostic test accuracy reviews Rücker G1 , Steinhauser S1 , Schumacher M1 1 Medical Center, University of Freiburg, Germany Background: In diagnostic test accuracy (DTA) reviews, it is often assumed that each study reports only one pair of specificity (Sp) and sensitivity (Se). The established bivariate model considers the joint distribution of Sp and Se across studies. However, in primary studies (Sp, Se) pairs are often reported for two or more cut-offs, and the cut-off values are reported as well. Objectives: To use this additional information for modelling the distributions of the underlying biomarker for diseased and non-diseased individuals and to determine an optimal cut-off. Methods: We assume that for some or all DTA studies in a meta-analysis a number of cut-offs is reported, each with corresponding estimates of Sp and Se. These provide direct information about the empirical cumulative distribution function (ecdf) for both groups of individuals. We propose a class of hierarchical models for the distribution including study as a fixed or random factor. These models lead to average ecdfs for both groups of individuals. As the difference between these is the Youden index as a function of the cut-off, an optimal cut-off can be found by maximising Cochrane Database Syst Rev Suppl 1–327 (2015) 161 this function. A summary receiver operating characteristic (ROC) curve is estimated based on the distributions. Results: The approach is demonstrated on three meta-analyses of alcohol screening, procalcitonin as a marker for sepsis and diagnosis of asthma. Conclusions: If there are a number of studies reporting at least two cut-offs with (Sp, Se) per study, we can determine an optimal cut-off and estimate a summary ROC curve based on all available information from the primary studies. LRO 11.6 Time trends in summary estimates from meta-analyses of diagnostic accuracy studies Cohen JF1 , Korevaar DA1 , Wang J1 , Leeflang MM1 , Bossuyt PM1 1 Department of Clinical Epidemiology, Biostatistics and Bioinformatics, Academic Medical Center, University of Amsterdam, The Netherlands Background: Summary estimates from meta-analyses of diagnostic accuracy studies can change over time, as evidence accumulates. Objectives: To evaluate trends in changes over time in summary estimates of diagnostic accuracy. Methods: We included 54 meta-analyses with at least five primary studies from 42 MEDLINE-indexed systematic reviews published in English between September 2011 and January 2012, reflecting 706 diagnostic accuracy studies, with 338,109 participants in total. Within each meta-analysis, we ranked studies by publication date and applied random-effects cumulative meta-analysis, to follow how summary estimates of diagnostic odds ratio (DOR), sensitivity and specificity would have evolved over time. Time trends were assessed by fitting a weighted linear regression model of the summary accuracy estimate against rank of publication. A negative slope indicates a decrease in summary accuracy estimates over time. Results: Slopes ranged from -0.80 to 0.56 for DOR, −0.44 to 0.37 for sensitivity and -0.37 to 0.36 for specificity. We found a significant time trend in at least one accuracy measure in 42 (78%, 95% confidence interval (CI) 65% to 87%) meta-analyses. Among a total of 93 significant time trends in DOR, sensitivity or specificity, 62 (67%, 95% CI 57% to 75%) were negative, indicating that summary estimates of accuracy tended to decrease over time. Conclusions: Changes in summary estimates are relatively frequent in meta-analyses of diagnostic accuracy studies, with estimates based on more studies typically reflecting weaker diagnostic performance. The results from early meta-analyses of diagnostic accuracy studies should be considered with caution before drawing conclusions about the performance of medical tests. Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Lunch rapid oral session 12 Decision-making: Policy and public health LRO 12.1 Evaluation of an organizational strategy for evidence-informed decision making Ward M1 , Bryant B1 , Dobbins M2 , Peirson L2 1 Peel Public Health, Canada; 2 McMaster University, Canada Background: Peel Public Health is one of Canada’s largest public health departments with 630 staff serving 1.3 million residents. In 2009 we set a 10-year strategy for evidence-informed decision making (EIDM). It involved development of a rapid review process, extensive work force development, consistent funding, supportive infrastructure, and highly visible leadership. Objectives: To conduct a midpoint evaluation of the strategy to assess the use and impact of research in program decisions. Methods: The strategy was evaluated in five ways. 1. Survey of senior leaders at three points about the impact of research in their division. 2. Assessment of the practice outcome from each of 40 rapid reviews. 3. Uptake of EIDM training by staff at all levels of the organization. 4. Assessment of the change in skills to find, appraise and apply research by an external researcher. 5. Case study at three points to assess the penetration of EIDM into the organization. Results: The main barriers to research use in 2009 were lack of access to full text articles, inconsistent ability to assess the quality of research, and no time. By 2014 we belonged to a library consortium and had two librarians, had trained 130 staff in EIDM, and had mentored 40 specialist-manager teams in the rapid review process. Skills in finding, appraising and applying research were tested and had improved significantly. Program changes arising from rapid reviews included five to stop a program, 20 to change a program, 12 to start a new program and nine to maintain the status quo. The case study revealed high visibility of the strategy for all staff, but minimal penetration to front line. Conclusions: Creating a culture for research use can be achieved by an organization. It requires strong and persistent senior leadership, investment in necessary infrastructure and funding, intensive staff training and mentorship, and an intentional change management strategy. Facilitators include close relationships with researchers, good quality research syntheses and the availability of well-tested training programs. Challenges include staff turnover, competing priorities, and organizational fatigue. Cochrane Database Syst Rev Suppl 1–327 (2015) 162 LRO 12.2 LRO 12.3 Developing a conceptual framework to support coverage decisions for vaccines adoption: a Delphi consensus process Influencing policy for promoting appropriate prescribing and medicines use: training key decision makers to use a database of systematic reviews González-Lorenzo M1 , Tirani M2 , Piatti A3 , Coppola L3 , Gramegna M3 , Auxilia F2 , Demicheli V4 , Banzi R5 , Bertelè V5 , Moja L1 1 Clinical Epidemiology Unit, IRCCS Galeazzi Orthopaedic Institute, Milan; Diparment of Biomedichal Sciences for Health, University of Milan, Italy; 2 Deparment of Biomedichal Sciences for Health, University of Milan, Italy; 3 Direzione Generale Salute, UO Governo della Prevenzione e Tutela Sanitaria, Regione Lombardia, Milan, Italy; 4 Servizio Regionale di Riferimento per l’Epidemiologia, SSEpi-SeREMI - CochraneVaccines Field, Azienda Sanitaria Locale ASL AL, Alessandria, Italy; 5 Laboratory of Regulatory Policies, IRCCS-Mario Negri Institute for Pharmacological Research, Milan, Italy Background: Decision makers must make choices about which vaccines to incorporate into the prevention programs of a National Health Service. A transparent, unbiased, and comprehensive framework based on evidence-based criteria is necessary to guide decision-making on vaccine adoption. We developed a potential framework to support vaccine coverage decisions based on the DECIDE Evidence to Decision framework. Objectives: This study aims to evaluate and validate a framework to support vaccine coverage decisions, and to explore alternative ways of presenting framework criteria to support policy decision-making. Methods: We adopted a Delphi consensus process to engage a group of experienced healthcare professionals. We constructed a questionnaire based on the six dimensions of the framework (burden of disease, vaccine characteristics and impact of immunisation programme, values and preferences, resource use, equity and feasibility). This resulted in 81 structured questions concerning the relevance of each proposed criterion. We requested participants to rate these criteria on a 9-point Likert scale ranging from 0 (not at all important) to 9 (extremely important). We then conducted a two-round Delphi process through Internet and a discussion group. Results: Fifty-nine participants from multidisciplinary areas were invited by email; 46 participants accepted the invitation. In the first round, 46 out of 81 criteria achieved consensus. Among those lacking consensus, the major concerns were related to criteria about mortality, morbidity, symptoms, and differences between local and global data. We presented results of the first round in a group discussion. In the second round, 48 criteria were assessed, and 41 achieved consensus. Conclusions: In order to optimize its application and effectiveness, a vaccine framework should consist of dimensions and criteria that are supported by evidence summaries; these can be presented in graphical or tabular displays. Additional research is required to test the framework for coverage decisions on one or more vaccines, in order to identify their critical elements, and their relative importance in influencing decision-making. Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Colquhoun H1 , Hellis E2 , Belanger D2 , Ens B2 , Hill S3 , Lowe D4 , Mayhew A5 , Taylor M6 , Grimshaw J7 1 University of Toronto, Canada; 2 Canadian Agency for Drugs and Technologies in Health, Canada; 3 Co-ordinating Editor, Cochrane Consumers and Communication Review Group/La Trobe University, Australia; 4 La Trobe University, Australia; 5 Ottawa Hospital Research Institute, Canada; 6 Australian Catholic University, Australia; 7 Director Cochrane Canada/Ottawa Hospital Research Institute/University of Ottawa/, Canada Background: Health systems globally promote appropriate prescribing by health care providers and safe and effective medicines use by consumers. Rx for Change (a publicly available database developed by the Cochrane Effective Practice & Care, and Consumers & Communication Review Groups) provides access to systematic reviews regarding best practices for prescribing and using medicines. We developed a training program, with knowledge user (KU) input, to promote its wider use. Objectives: To describe the development, delivery, and evaluation of an Rx for Change training program for five medicine-focused organizations in Canada and Australia. Methods: Development: Interviews with KUs were conducted in each organization. A directed content analysis of interview transcripts was performed and used to develop generic and tailored training content. Delivery: Workshops were offered to all organizations. Evaluation: A survey measuring attitudes, skills, and intentions for using Rx for Change and evidence for decision-making was conducted pre-, post-, and three months after training completion. Results: Development: Seventeen KUs participated in interviews. Themes for training content included the scope of, navigation and strategies for using Rx for Change (generic content) and practical examples on incorporating evidence within their workplace context (tailored content). Delivery: The program included an information video, a 60-minute face-to-face workshop with didactic, hands-on and interactive components, and two post-training reminders. Forty-eight people attended one of six workshops. Evaluation: ANOVA indicated statistically significant improvements in: skills for using evidence in policy decision-making F(2,117) = 8.034, P = 0.001; skills in using Rx for Change, F(2,117) = 88.58, P = 0.000; and attitudes towards Rx for Change F(2,117) = 7.079, P= 0.001. Confidence in skills for using evidence and Rx for Change was maintained three months after training (P ¡0.007; P ¡0.000). Conclusions: Sustained improvements in skills for using evidence in policy decision-making were reported by participants in a training program developed with KU involvement. Cochrane Database Syst Rev Suppl 1–327 (2015) 163 LRO 12.4 LRO 12.5 Does presenting the findings of a review change public health decision-makers’ view on an intervention? Translating health research to policy: taking forward evidence-informed decision-making in Nepal Baker P1 , Francis D1 , Demant D2 1 Public Health, School of Public Health & Social Work, Queensland University of Technology, Australia; 2 School of Public Health & Social Work, Queensland University of Technology, Australia Karki K1 , Subedi M2 , Karki A1 , Aryal K1 , Stewart M3 , Rana S4 , Dhimal M1 1 Nepal Health Research Council, Nepal; 2 Patan Academy of Health Sciences, Nepal; 3 Ministry of Health and Population, Nepal; 4 London Borough of Merton, United Kingdom Background: Knowledge translation and exchange strategies are required to support the utilisation of Cochrane Reviews in decision making. Objectives: To analyse the views of public health practitioners and whether they change after participating in a workshop communicating the findings of a recently updated relevant review. Methods: Audience response technology (clickers) were used to enhance engagement and identify participants’ views on the topic. Participants from three Canadian public health units were polled before and after the presentation. The polling included an assessment of agreement with the statement ’Community-wide interventions effectively increase population levels of physical activity’. The review found the interventions rarely provided an effect. Results: Across the three units, 132 participants used clickers. Initially, 77% of respondents stated they had confidence in identifying an evidence-based program or service. Prior to presentation 3% strongly agreed, and 36% agreed with the statement ‘Community-wide interventions effectively increase population levels of physical activity’. Post presentation, the agreement with the statement reduced to 1.5% strongly agreeing and 3% agreeing. Eighteen per cent voted neutral and the remaining 77.5% disagreed. A small proportion (6%) of participants increased their belief in the intervention (by 1 level on a 5 point scale), and 26% remained unchanged in their view. However, a large proportion of the participants (68%) reduced their belief by 1 or more level (Wilcoxon signed ranks test: Z= -6.850, P value < 0.001). Those who had prior belief in community-wide interventions had the greatest likelihood for change. Sixty per cent agreed with the presentation findings, 30.5% remained neutral and 9.5% disagreed. Conclusions: Presenting findings of systematic reviews to stakeholders has the potential to modify participant’s views. The use of clickers can identify these changes and stimulate discussion. Further research is necessary to identify whether long-term policy change occurs from the presentation. Background and Objectives: Knowledge derived from research and experience is of limited value if not put into practice. Currently, attention is increasingly focused on innovative ways to introduce and scale-up evidence-informed practices in the health sector. It is estimated that only 14% of findings actually make the transition from research to practice within two decades of their discovery. This research is designed to understand factors that facilitate or hinder the translation of evidence into policy in Nepal, and to identify ways forward for improvement and developing mechanisms to link health research to policy. Methods: Key informant interviews will be conducted with researchers and policy makers using separate interview guides. A minimum of 15 interviews each with researchers and policy-makers will be undertaken. After preliminary analysis of interviews, a separate mini-workshop will be conducted with each group to explore issues raised further. Results: Our study is currently in progress and we expect to have our results to present at the Colloquium. We expect this research to identify key gaps on both sides; researchers, who are the evidence-creators and policy-makers, who are the users of evidence. We expect the findings to guide further actions of evidence synthesis and knowledge translation in Nepal and similar low- and middle-income countries, and ultimately bridge the knowledge gap that exists between researchers and policy makers. This research is well accepted by the researchers and policy makers during the interviews. Conclusions: Regular assessment of gaps in knowledge translation and the development of a mechanism of communication between researchers and policy-makers will help to translate research into evidence-informed decision-making. LRO 12.6 A methodological approach for binary data to determine the extent of treatment effect in the early benefit assessment of drugs Bender R1 , Skipka G1 , Wieseler B1 , Kaiser T1 , Thomas S1 , Windeler J1 , Lange S1 1 IQWiG, Germany Background: At the beginning of 2011 the early benefit assessment of new drugs was introduced in Germany with the Act on the Reform of the Market for Medicinal Products (AMNOG). The Federal Joint Committee (G-BA) generally Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Cochrane Database Syst Rev Suppl 1–327 (2015) 164 commissions the Institute for Quality and Efficiency in Health Care (IQWiG) with this type of assessment, which examines whether a new drug shows an added benefit over the current standard therapy on the basis of a dossier submitted by the pharmaceutical company. In this benefit assessment the costs of the therapies are not taken into account. In the case of an added benefit, the law requires this benefit to be classified as minor, considerable, or major. Objectives: To describe and discuss the approach for assessing the extent of added benefit for binary data developed by IQWiG. Methods: Based on the underlying law, outcomes are grouped according to their relevance as follows: (1) mortality, (2) serious (or severe) symptoms (or late complications); serious (or severe) adverse events, health-related quality of life, (3) non-serious (or non-severe) symptoms (or late complications), non-serious (or non-severe) adverse event. Thresholds for confidence intervals for the relative risk are derived by means Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. of meta-analytic power considerations for shifted hypotheses based on suggested effects for each of the categories of added benefit for each outcome group. Results: For the treatment effect to be classified as a minor, considerable or major added benefit, the methodological approach stipulates that the (two-sided) 95% confidence interval of the effect must exceed a specified distance to the zero effect. Conclusions: The proposed approach provides a robust, transparent and thus predictable foundation to determine minor, considerable and major treatment effects on binary outcomes in the early benefit assessment of new drugs in Germany. After a decision on the added benefit of a new drug by G-BA (Gemeinsame Bundesausschuss), the classification of added benefit is used to inform pricing negotiations between the umbrella organization of statutory health insurance and the pharmaceutical companies. Cochrane Database Syst Rev Suppl 1–327 (2015) 165 Poster Session 1 Impact of Cochrane reviews P1-003 Analysing the Altmetric scores of articles from the Cochrane Database of Systematic Reviews Stewart G1 , Ruotsalainen J2 1 Wiley, United Kingdom; 2 Cochrane Occupational Safety and Health Review Group, Finland Background: In 2014, Wiley partnered with Altmetric; a company that compiles article-level metrics. Altmetric provides an alternative to the more traditional ways of measuring the impact of scholarly work, by tracking data derived from online activity and discussions about individual scholarly papers collectively known by Altmetric as ‘mentions’. Cochrane Reviews are given an Altmetric score based on the number of mentions they receive. Wiley use traditional metrics to provide an annual ‘Impact and usage’ report to every Cochrane Review Group (CRG) with the aim of assisting groups in showing the impact of their published reviews. The report does not currently include Altmetric data and the use of Altmetric data within Cochrane is not widespread. Objectives: To analyse Altmetric data and establish whether more widespread use of the data would be beneficial for CRGs and Cochrane authors. Methods: An analysis of the number of mentions received by all published Cochrane articles and a more detailed analysis of mentions received by articles published by the Cochrane Occupational Safety and Health (OSH) Group will be undertaken. The online submission system, Archie, will be used to export the publication details of all published Cochrane articles. The Altmetric Explorer platform will be used to export the number of mentions associated with all published Cochrane articles and to analyse the nature of mentions received by articles from the OSH group. Results: To date, 6040 Cochrane articles have received an Altmetric score; 34 of these articles were published by the OSH group. The results will show how CRGs compare by score and weighting of mentions. Activity associated with articles published by the OSH group will be analysed additionally by demographics and source. Conclusions: Conclusions will be drawn and presented from the analysis. P1.004 Identifying Cochrane citation classics Stewart G1 , Uthman O2 , Pentesco-Gilbert D1 1 Wiley, United Kingdom; 2 University of Warwick, United Kingdom Background: The concept of ‘citation classics’ was first mentioned in a 1977 essay by the creator of the Impact Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Factor, Eugene Garfield (Garfield 1977). The first group of citation classics contained the 500 papers most cited from 1961 to 1975. In October 2013, a paper published in PLoS ONE identified and assessed a group of the 100 most cited articles in systematic reviews and meta-analyses from 1977 to 2008 (Uthman 2013). The 100 most cited articles did not contain a Cochrane Review despite Cochrane Reviews being internationally recognised as the highest standard in evidence-based health care. Objectives: To identify and examine the characteristics of a group of citation classics from the Cochrane Database of Systematic Reviews. A secondary aim is to investigate how citation classics from the Cochrane Database of Systematic Reviews compare with citation classics from other systematic reviews and meta-analyses published in the same time period. Methods: The Web of Science Core Collection of Thomson ISI will be searched to identify the most cited Cochrane Reviews from 2005 to 2014. Data on the year of publication, associated keywords, average number of cites per year, the country and institution of the corresponding author and the Cochrane Review Group (CRG) responsible for producing the review will be gathered. Results: Initial investigations have found that 7068 Cochrane Reviews have been indexed in the Web of Science Core Collection from 2005 to 2014. The highest cited Cochrane Review in this time period has been cited 452 times. The 100th highest cited Cochrane Review has been cited 89 times. The results of the data analysis will be made available at the meeting. Conclusions: By identifying Cochrane citation classics it is hoped that opportunities for future engagement in topic areas and with institutions and individuals will be discovered. References Garfield E. Introducing citation classics –human side of scientific reports. Curr Contents 1977; 1: 5–7 Uthman OA, Okwundu CI, Wiysonge CS, Young T, Clarke A. Citation classics in systematic reviews and meta-analyses: who wrote the top 100 most cited articles? PLOS ONE 2013: 8:10 P1.006 Open access: collaboration between NIHR and Wiley Osborne R1 , Bailey S1 1 National Institute for Health Research, United Kingdom Background: The UK National Institute for Health Research (NIHR) provides GBP 1.4 million a year direct funding to Cochrane Reviews through the Cochrane Programme Grants Scheme, established to provide high-quality systematic reviews of direct benefit to users of the National Health Service. Objectives: To comply with NIHR Open Access policy, the NIHR reached an agreement with Cochrane and Wiley in 2013 to deposit all Cochrane Reviews produced from the Cochrane Programme Grants to Europe PubMed Cochrane Database Syst Rev Suppl 1–327 (2015) 166 Central (PMC) 12 months after publication. Methods: The submission of each Cochrane Review to Europe PMC is reviewed and approved by the Cochrane Editorial Unit, as detailed in the Cochrane Editorial and Publishing Policy Resource (EPPR). Upon publication of a review on the Cochrane Library, Wiley supply the standard review PDF to NIHR for display on the NIHR Journals Library website. Upon expiration of the 12-month embargo, the PDF is taken down from the website, as it is free to access in The Cochrane Library from that point onwards. Results: All Cochrane Programme Grant Reviews published on or after 1 February 2013 meet the Wiley ‘green’ open access policy, and have been placed on the NIHR Journals Library website under a dedicated collection: http://www.journalslibrary. nihr.ac.uk/nihr-research/cochrane-programme-grants. After the initial 12 months they are free to access for all via the Cochrane Library, and can be included in an institutional repository or any repository mandated by the author’s funder, such as PMC, 12 months after publication. Conclusions: ‘NIHR investment in systematic reviews means our health and care services have access to the best possible evidence to inform decisions and choices’ (Professor Dame Sally Davies, CMO, England). Our collaboration with Wiley has made alignment with the NIHR Open Access Policy a reality. The Department of Health and NIHR agree with the principal of Open Access to the outputs of its research, recognising it can offer both social and economic benefits, as well as aiding the development of new research and stimulating wider economic growth of the UK economy. studied were likely to be beneficial, and they all recommended additional research. Also, two of the reviews concluded that the interventions were not helpful, one of which did not recommend further studies and one review recommended additional studies. In total, 38 of the reviews reported that the evidence did not support either benefit or harm, two did not recommend further studies and 36 recommended additional studies. Therefore, 81 of the total of 84 completed SRs recommended further research (Fig 1). Conclusions: Forty-four Cochrane SRs investigating efficacy of CBT for prevention and treatment of mental disorders concluded that the interventions studied were likely to be beneficial, and they all recommended additional research. Thirty-eight reviews reported that the evidence did not support either benefit or harm. On the basis of these results we recommend the elaboration of high-quality primary studies in CBT in active collaboration and consultation with International Societies of CBT and Academy. This will represent a major component of methodological advance in dissemination of Evidence Based Psychology. P1.007 Mapping the Cochrane Library for decision making in Cognitive Behavior Therapy (CBT) for prevention and treatment of mental disorders 1 2 2 Melnik T , Neufeld CB , Ferreira IMF , Atallah AN 1 Brazilian Cochrane Center, Brazil; 2 University of Psychology Ribeirão Preto - FFCLRP, Brazil Background: Cognitive behavior therapy (CBT) is based on the main assumption that how we perceive situations influences how we feel emotionally and how we act in different situations. CBT helps people identify their dysfunctional thoughts and evaluate how realistic those thoughts are. The focus is on learning how to change distorted thinking. The emphasis is also consistently on solving problems and initiating behavioral change. Objectives: We evaluated the conclusions from Cochrane Systematic Reviews (SRs) of randomized controlled trials in terms of their recommendations for clinical practice and research for CBT for prevention and treatment of mental disorders. Methods: In our cross-sectional study of systematic reviews published in the Cochrane Library, we analyzed all the completed systematic reviews published about the efficacy and effectiveness of CBT in the prevention and treatment of mental disorders. Results: We analyzed 84 completed SRs. Of these, 44 (52.4%) concluded that the interventions Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Mixed methods/implementation 1 P1.009 A novel methodological approach to synthesising the findings from quantitative and qualitative systematic reviews Moore D1 , Gwernan-Jones R1 , Ford T1 , Garside R1 , Stein K1 , Rogers M1 , Thompson Coon J1 1 University of Exeter Medical School, United Kingdom Background: There are few established methods for the integration of findings from quantitative and qualitative systematic reviews of interventions. This paper describes a method of overarching synthesis that we developed to bring together the findings from four systematic reviews about the effectiveness and experience of school-based interventions for attention-deficit/hyperactivity disorder (ADHD). Methods: We approached the synthesis from two perspectives: 1. we brought together the findings of two qualitative syntheses using an inductive approach to Cochrane Database Syst Rev Suppl 1–327 (2015) 167 identify the contextual elements that might influence the effectiveness of interventions; 2. we worked deductively from quantitative findings about the effectiveness of interventions and moderators of effect, to identify findings in our other reviews that offered hypotheses about the relationships between possible moderators and effectiveness. Results: Findings from the inductive approach revealed a hierarchy of contextual levels that influence the use of school-based interventions. Implications for the development and evaluation of school-based ADHD interventions were raised from the deductive findings which helped to further our understanding of potential moderators of effectiveness. Unlike existing methods of overarching synthesis, the methods employed did not only explain the findings of the effectiveness review, but rather developed a greater understanding of the use of ADHD interventions in school settings. Conclusion: The addition of an overarching synthesis to this series of systematic reviews has allowed comparison across all reviews, and brought together a breadth of evidence regarding the use of ADHD interventions in school settings. The identified links between reviews are valuable because of their relevance to future intervention programme design and evaluation. P1.010 Appropriate use of prescription opioids: how to do a systematic review of strategies Irvin E1 , Furlan A1 , Van Eerd D2 , Munhall C1 , Kim J3 , Danak S1 1 Cochrane Back Review Group, Canada; 2 Insitute for Work and Health, Canada; 3 Cancer Care Ontario, Canada Background: The World Health Organization (WHO) estimates that worldwide 69 000 people die from opioid overdose each year; an estimated 15 million people suffer from opioid dependence (i.e. an addiction to opioids). Objectives: To determine existing strategies, frameworks, collaborative networks, and materials aimed at improving the appropriate use of prescription opioids, and/or reducing the misuse, abuse and diversion of these drugs. Methods: We conducted a mixed methods review. We searched MEDLINE, EMBASE, PsycINFO, and CINAHL, and identified grey literature by searching websites for the following groups: regulatory authorities for health-related professionals, government, public health and health promotion agencies, prevention and treatment organizations, workers’ compensation boards, private insurance companies, and law enforcement agencies. There were no language restrictions. Our interventions of interest were the types of strategies developed to improve the appropriate use of opioids. Our primary outcomes were: appropriate opioid use for pain, reduction of misuse, abuse, and/or addiction, overdose, diversion and/or crime. We engaged stakeholders from law enforcement, government, public health, and vulnerable groups. Results: We found a total of 3191 references from the database searches and grey literature. We Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. included 109 references: 26 empirical studies, 53 descriptive studies, and 30 reviews. Some studies only described the strategies, but did not have any empirical evidence to support the implementation of the strategy. We classified the strategies into 14 groups: educational, promotional, campaign, framework, interactions, networking, community, forum, actions, regulations/law enforcement, prescription monitoring, programs, take-home naloxone, and others. Conclusions: This mixed method review found 14 types of strategies that were relevant and could be disseminated to consumers. The empirical evidence may be examined to provide the effectiveness of some of these strategies on improving the appropriate use of prescription opioids. Outputs from this review included proposed policy, safety kits, and public awareness campaigns. P1.011 Evidence mapping - do systematic reviews report the inclusion of women and sex as a moderator of intervention effects? Williams JW1 , Duan-Porter W1 , Allen LaPointe N1 , Goldstein KM2 , Hughes JM3 , McDuffie JR2 , Klap R4 , Gierisch JM1 1 Duke University, USA; 2 Durham VA Medical Center, USA; 3 University of North Carolina at Chapel Hill, USA; 4 Greater LA VA Medical Center, USA Background: Delivering effective healthcare for women is a priority for policymakers and requires evidence about effective interventions. Objectives: To evaluate reporting on the inclusion of women and evaluation of sex-specific intervention effects in systematic reviews (SR). Methods: We engaged stakeholders to prioritize conditions and interventions. We searched PubMED and the Cochrane Library for English-language SR published since 2009, that addressed a medication, behavioral, quality improvement (QI), or exercise intervention for depression, diabetes mellitus, fibromyalgia, chronic low back or knee pain. We summarized data on the inclusion of women, intervention effects and sex effects as a moderator using descriptive statistics. Results are reported for depression; evaluation of the other conditions is in progress. Results: We screened 629 citations for depression, and 79 SR were eligible. The most commonly evaluated conditions were: major depression (n = 54 studies), minor depression (n = 22), dysthymia (n = 21), and treatment-resistant depression (n = 5). Evaluated interventions included: psychotherapy (n = 47), antidepressants (n = 34), relapse prevention strategies (n = 6), guided self-help (n = 4), exercise (n = 1) and QI (n = 1). The proportion of women enrolled in the primary studies was reported in 27 SR (34%) and in this subset, women constituted ≥ 50% of the sample in all pooled estimates. Sex-specific effects were evaluated in 14 SR (18%), using meta-regression (n = 7), patient level meta-analyses (n = 3), subgroup analyses (n = 2), or qualitative synthesis (n = 2). Four SR reported no sex effects for: therapy vs antidepressant (AD), therapy and AD vs AD alone, self-help, and QI. Cognitive behavioral therapy and antidepressants (i.e. SSRI and venlafaxine) were found Cochrane Database Syst Rev Suppl 1–327 (2015) 168 to be more effective for women than men. Sexual adverse effects were more common in men than women treated with paroxetine. Conclusions: The minority of SR described the inclusion of women or evaluated sex as a moderator of treatment effect. Some depression interventions appear more effective for women. However, moderator effects were often evaluated using suboptimal methods. P1.012 Reporting of intervention fidelity in systematic reviews of asthma education Ospina M1 , Villa-Roel C2 , Rowe BH2 1 Alberta Health Services, Canada; 2 University of Alberta, Canada Background: Intervention fidelity is the degree to which an intervention is delivered as intended. Recent evidence indicates that intervention fidelity should be part of the appraisal of randomized controlled trials (RCTs) of non-pharmacological interventions. It is not known how systematic reviews (SRs) of non-pharmacological RCTs in asthma education evaluate intervention fidelity and which specific fidelity dimensions have been reported. Objectives: To describe the reporting of intervention fidelity in SRs of RCTs assessing the effectiveness of asthma education. Methods: Relevant SRs of the effectiveness of asthma education in children and adult populations were identified through searches in seven electronic databases. Two independent reviewers assessed eligibility of SRs and completed data extraction. The Treatment Fidelity Assessment Grid was used to evaluate the reporting of five dimensions of treatment fidelity in the SRs: (1) fidelity to theory; (2) provider training; (3) treatment implementation; (4) treatment receipt; and (5) treatment enactment. Results were summarized descriptively. Results: From 5218 citations, 24 SRs of asthma education (10 Cochrane versus 14 non-Cochrane reviews) were included. The majority of reviews (n = 20) assessed at least one aspect of treatment fidelity of individual RCTs. The dimensions of treatment fidelity most frequently assessed in SRs were treatment implementation (n = 17) and provider training (n = 11); treatment receipt (n = 6), theoretical fidelity (n = 4) and treatment enactment (n = 4) were less frequently reported. No statistical associations between these domains and SRs characteristics (type, quality and direction of results) were found. Conclusions: Current SRs of asthma education reveal inconsistent reports of intervention fidelity in RCTs. This limits the translation of results into clinical practice and may influence the interpretation of effectiveness. Using standardized models of fidelity assessment in SRs will ensure that key features of non-pharmacological interventions are understood, appraised appropriately, and examined in relation with therapeutic change and outcome results. Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. P1.013 Methodological strategies for the synthesis of complex interventions –oral health education programmes for nursing staff and residents Albrecht M1 , Kupfer R1 , Reißmann DR2 , Mühlhauser I1 , Köpke S3 1 MIN Faculty, Unit of Health Sciences and Education, University of Hamburg, Germany; 2 Department of Prosthetic Dentistry, University Medical Center Hamburg-Eppendorf, Germany; 3 Nursing Research Unit, Institute of Social Medicine, University of Lübeck, Germany Background: Methodological challenges regarding the synthesis of complex interventions have been discussed extensively. Meta-analyses of randomized controlled trials (RCTs) do not provide sufficient information concerning the implementation of complex interventions such as education programmes. Contextual and component-related factors are indispensable for review users. So far this information is rarely provided in Cochrane Reviews. Objectives: Produce a description of intervention components and context factors in a Cochrane Review on the effectiveness of oral health promotion education in nursing homes. Methods: The systematic review was prepared in accordance with the Cochrane Handbook. Data retrieval on research processes of the development, piloting and evaluation of the complex interventions including description of single components and context factors based on the reporting guideline ‘Criteria for Reporting the Development and Evaluation of Complex Interventions in healthcare’ (CReDECI2). In addition to RCTs, other publications on development, piloting and process evaluations were identified and included through additional systematic searches, author requests and reference tracking. Results: All included studies (2 RCTs, 7 cluster-RCTs) were complex interventions using more than one active component. Results for dental health indicate that educational interventions may have positive short-term effects on dental plaque, denture plaque, and inflammation of oral mucosa (moderate quality evidence). Included studies reported insufficient information on development and evaluation processes of complex interventions. For example, only two studies reported on reasons for the selection of intervention components and no study included information on relevant context factors (Figure 1). Conclusions: Oral health educational interventions seem to positively influence residents’ dental health in the short term. Conclusions about effective components cannot be made due to insufficient reporting. Facilitators, barriers and resources of the different interventions are unknown, but are crucial for decisions on implementation. Cochrane Database Syst Rev Suppl 1–327 (2015) 169 P1.014 Practical tools and guidance on conducting systematic reviews of complex interventions Viswanathan M1 , Guise J2 , Chang C3 , Butler M4 , Kondo K2 , Dickinson C2 , Arbues F2 , Motu’apuaka M2 , Pigott T5 , Tugwell P6 1 RTI-UNC Evidence-Based Practice Center, RTI International, USA; 2 Oregon Health Sciences University, USA; 3 Agency for Healthcare Research and Quality, USA; 4 Minnesota Evidence-Bsed Practice Center, University of Minnesota, USA; 5 Loyola University, Chicago, USA; 6 University of Ottawa, Canada Background: 1. The Agency for Healthcare Research and Quality (AHRQ), under the aegis of the United States Department of Health and Human Services, supports the conduct of systematic reviews (SRs) on a wide array of clinical topics. With a rapidly changing healthcare environment, AHRQ’s efforts to conduct SRs of complex interventions are taking on increasing importance. AHRQ is supporting a meeting of experts in complex interventions to identify practical tools and guidance for scoping reviews, creating appropriate analytic frameworks, and analyzing data (June 2015, Washington DC). Objectives: To create practical tools for and guidance on conducting systematic reviews of complex interventions. Methods: Guidance will be issued Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. around best practice and applications of new methodological approaches for analysis of complex interventions, based on a group consensus process of international experts in evidence reviews, research, and healthcare implementation of complex interventions. Results: SRs of complex interventions can vary widely in purpose, data availability and heterogeneity, and end-user expectations. Questions from end-users may range from exploratory to confirmatory. Underlying sources of complexity can arise from context, intervention, implementation, and participant responses. The analytic framework chosen to answer systematic review questions will need to account for end-user needs, sources of complexity, and causal or associative links. Analytic approaches can vary in data requirements or underlying causal assumptions. We summarize the guidance developed through multiple workgroups. Specifically, we touch upon context, frameworks, and tools for complex intervention; address the range of questions posed in SRs; and the analytic approaches available to answer these questions. We offer a decision framework to select the analytic approach best suited to the context of the systematic review and note data requirements, expertise needed, statistical power, and computing resources. Conclusions: The resulting guidance will aid a global audience of systematic reviewers of complex interventions in improving the utility and rigor of their reviews. Cochrane Database Syst Rev Suppl 1–327 (2015) 170 Education and training P1.016 P1.015 Iranian biomedical researchers and their contribution to developing systematic reviews Experience of capacity building for evidence-based child health Agarwal A1 , Singh M2 , Jaiswal N1 , Thumburu KK1 , Mathew JL2 , Malhotra S3 , Shafiq N3 , Dutta S2 , Chadha N4 1 ICMR Advanced Centre for Evidence-Based Child Health, Postgraduate Institute of Medical Education and Research, Chandigarh, India; 2 Advanced Pediatric Centre, Postgraduate Institute of Medical Education and Research, Chandigarh, India; 3 Department of Pharmacology, Postgraduate Institute of Medical Education and Research, Chandigarh, India; 4 Dr.Tulsi Das Library, Postgraduate Institute of Medical Education and Research, Chandigarh, India Background: There is a growing need for evidence to inform policy and practice in India. The Post Graduate Institute of Medical Education and Research, Chandigarh under the aegis of the Indian Council of Medical Research (ICMR), New Delhi, established an advanced center for evidence-based child health. This center conducted workshops and short courses to inform healthcare providers and researchers about the principles of systematic reviews (SRs). Objectives: To build capacity, conduct SRs and promote practise and training of evidence-based health care in children. Methods: Twelve short courses on ‘How to practise evidence-based child health’, six short courses on ‘How to teach evidence-based medicine’ and nine workshops on ‘Protocol development’ were conducted in Chandigarh and two satellite centers, two each in north and north-east India, by the tutors from the ICMR advanced centre for evidence-based child health. We followed the module developed by the Royal College of Pediatrics and Child Health. We conducted pre-testing and post-testing with similar questions during each course to evaluate the knowledge of evidence-based child health. Short courses included lectures followed by small group interactive sessions on critical appraisal of randomized controlled trials, diagnostic test accuracy and conduct of SRs. Results: Six-hundred and fifty participants from different parts of India were educated and introduced to Cochrane and SRs. More than half of the participants were not aware of Cochrane and SRs before the course. More than 50% indicated interest in further training in SRs. Most participants preferred small group interactive sessions. There was a significant increase in post-test marks compared to pre-test. Conclusions: Capacity-building workshops and short courses in India are avenues for increasing awareness and contributions to Cochrane and SRs. There is a growing need to continue to train individuals and develop programmes to support the use of evidence for policy making and clinical practice in India. Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Mesgarpour B1 , Arabzadeh S2 , Kabiri P3 , Akhondzadeh S4 , Malekzadeh R5 1 National Institute for Medical Research Development (NIMAD), Iran; 2 Deputy of Research and Technology, Ministry of Health and Medical Education, Iran; 3 Department of Epidemiology & Biostatistics, School of Public Health, Tehran University of Medical Sciences, Iran; 4 Psychiatric Research Centre, Roozbeh Hospital, Tehran University of Medical Sciences, Iran; 5 Digestive Disease Research Institute, Tehran University of Medical Sciences, Iran Background: Appropriate training and the availability of key resources play a pivotal role in producing a systematic review. Over the past 10 years, the Ministry of Health and Medical Education in Iran has provided access to research databases and journals and has conducted several workshops on how to write systematic reviews for researchers in medical universities. Objectives: This study aimed to identify the systematic reviews/meta-analysis published by Iranian biomedical researchers, establish the overall trend and describe characteristics of such publications. Methods: A sensitive search strategy was designed based on authors’ affiliation and the key words ‘systematic AND review’ or ‘meta analysis’. The citation databases of Scopus and Web of Science were searched from 1990 to March 2015. We also searched for the Cochrane Database of Systematic Reviews as a source of publication. We identified systematic review/meta-analysis that affiliated by at least one author from Iran. Two independent reviewers screened and evaluated retrieved records to select relevant studies. We excluded study protocols. Results: Our search resulted in the retrieval of 2596 records (1735 after removing duplicates), of which 912 records (850 papers and 62 conference proceedings) were judged to be relevant. The number of papers increased from one in 1998 to 219 in 2014. The majority of papers were published in English (98%) and in non-Iranian journals (71%) including 43 papers in the Cochrane Database of Systematic Reviews. Out of 850 papers, 155 (18.2%) were conducted to investigate a national issue. The various domains of reviews were detected, and included treatment/prevention (37.2%), epidemiology and prevalence (14.9%), risk/harm assessment (10.2%), prognostic (9.9%), policy making (7.7%), public health (7.6%), diagnostic (6.1%) and gene disease association (4.8%). Conclusions: This study introduces the fast growth of scientific communications of Iranian biomedical researchers in developing systematic reviews. Further study on the quality assessment of these publications is warranted to identify potential gaps and educational demands for Iranian researchers. Cochrane Database Syst Rev Suppl 1–327 (2015) 171 P1.017 P1.019 Two years’ experience for training appraisers of guidelines using an AGREE II scoring guide in Korea Development of the first Academic League of Evidence-based Medicine in Brazil. A partnership between the Brazilian Cochrane Centre and Universidade Federal de São Paulo Shin E1 , Chang S2 , Kim D3 , Jang J4 , Yeon J1 , Lee Y5 1 Korean Academy of Medical Sciences, South Korea; 2 Department of Urology, Kyung Hee University School of Medicine, South Korea; 3 Department of Diagnostic Radiology, Yon Sei University School of Medicine, South Korea; 4 Korean Academy of Medical Sciences, Korea South; 5 Department of Legal Medicine, Seoul National University College of Medicine, South Korea Background: An effective education program had been developed in Korea since 2013 to train qualified appraisers of guidelines. To alleviate discrepancies among appraisers (multidisciplinary physicians), a scoring guide that consists of 92 guides for anchor points 1, 3, 5, and 7 in 23 items of AGREE II instrument has been developed and applied by the Executive Committee for clinical practice guidelines (CPGs), the Korean Academy of Medical Sciences in Korea. Objectives: To train and examine AGREE II scoring guide for time-saving and increasing reliability for the assessment of quality of CPGs among appraisers. Methods: Three hours of an education program was provided to 29 physicians in 2013 and 20 physicians in 2014. There were three different training subjects: understanding the AGREE II instrument and a scoring guide; learning how to apply the scoring guide; and how to evaluate each of the 23 items and assign a score. The guideline for prevention and treatment of metabolic syndrome in primary care and the guideline for the diagnosis and treatment of Helicobacter pylori infection in Korea, 2013 Revised Edition, were assessed using a Korean AGREE II scoring guide in 2013 and 2014. Each item was rated on a 7-point scale by participants. Disagreement was defined by more than four differences in score among appraisers. Results: Item 5 in domain 2 (stakeholder involvement) showed the highest disagreement (27.6%, 8 of 29), and five items (items 1, 4, 6, 7 and 15) showed no disagreement in 2013. Items18 and 19 in domain 5 (applicability) showed high disagreement (35% and 40.0%), and eight items (item 1, 2, 6, 7, 8, 10, 12 and 17) showed no disagreement in 2014. Conclusions: We demonstrated that the Korean AGREE II scoring guide was a useful tool to train appraisers of CPGs; it provides a capable time-saving method to teach how to rate with a low disagreement rate. Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Daou J1 , Atallah A1 , Barroso G1 , Costa I1 , Paroni R1 , Freitas R1 , Vitor S1 , Viel J1 , Klepp A1 , Baldan A1 , Martins J1 , de Mello S1 , Gomes D1 , Fonseca E1 , Gomes V1 , Protetti N1 , Hungria F1 , Porfirio G1 , Gois A1 , Riera R1 1 Universidade Federal de São Paulo, Brazil Background: The relationship between Universidade Federal de São Paulo and the Brazilian Cochrane Centre (BBC), which has some members as professors in the Institution, made possible the development of the first Academic League of Evidence-Based Medicine (EBML) in Brazil. Objectives: The objective of the EBML is to allow the student make a link between issues of great medical relevance and the evidence in the literature about those issues, leading to scientific knowledge and critical appraisal as well. These two components will guide the student’s best clinical practice in the future. Methods: The League’s activities include a weekly meeting for discussions and monthly supervised shifts in the emergency service. The topics for discussion include critical evaluation of literature, how to write scientific papers and how and where find the best evidence. Results: Eighteen months after the League’s creation (35 meetings), the following activities were accomplished: translation of abstracts of some Cochrane Systematic Reviews, a partnership with Students 4 Best Evidence, development of five papers (methods issues and case reports) for publication, an eight-hour course of evidence-based medicine (EBM) for the local academic community, creation of a social network page to disseminate EBM and participation of EBML members in the systematic review workshops from BCC. Conclusions: The EBML seemed to be a useful tool for dissemination of EBM among undergraduate students, for the creation of a critical scientific thought. It allowed the students to have more practice both medical and scientific, allowing them to enter their internships safer in their actions and crucially in the process of health decision-making with concepts of EBM. Cochrane Database Syst Rev Suppl 1–327 (2015) 172 P1.020 P1.022 Development and evaluation of doktormitSDM, an approach to improving SDM in daily clinical practice –a continuum of increasing evidence Multi-professional teaching of health research methodology in undergraduate programs. The case of Universidad de La Frontera in Chile Kasper J1 , Liethmann K2 , Klemperer D3 , Rumpsfeld M4 , Geiger F5 1 Department of Health and Caring Sciences,Faculty of Health Sciences, University of Tromsø, Norway; 2 Unit of Health Sciences and Education, University of Hamburg, Germany; 3 Ostbayerische Technische Hochschule Regensburg, Germany; 4 Division Internal Medicine in the UNN, Norway; 5 University Medical Center Schleswig-Holstein, Dept. of Pediatrics, Kiel, Germany Seron P1 , Orellana JJ1 , Baeza B1 , Velazquez M1 1 Universidad de La Frontera, Chile Background: Proven training of health professionals’ in shared decision making (SDM) competencies are rare. Sound evaluation requires consideration of such measures’ complex character. Objectives: We aim to structure our reporting on a series of studies to develop short in situ training for clinicians according to the complex interventions framework. Methods: We started with a comprehensive literature search. Moreover, theory and conceptual knowledge were worked up with regard to didactic requirements. Core components such as measurement-based video feedback or a video tutorial were piloted in scientific workshops. A pretest with 10 clinicians assessed patient involvement with the MAPPIN’SDM before and after the so called doktormitSDM training. Then, in a multicenter RCT (N = 40), SDMmass was used as primary endpoint, a compound measure for patient involvement. To allow for broad scale implementation, doktormitSDM didactic and contents were translated into an online format. The tutorial addresses a population of 12,000 practitioners counselling colon cancer screening decisions. We piloted the new format and newly developed components as an interactive video learning parkour. Results: The training has been shown to respond to a lack of research and meets specific needs by applying criteria of SDM and EBPI in a theory-based didactic concept. Components and entire version are proven feasible, time economic and motivating for clinicians. Promising results of the pre/post study informed the training’s revision and the RCT design. DoktormitSDM turned out efficient in changing communication quality of clinician-patient decision-making dyads. The RCT, however, still revealed practical difficulties in delivery to trainees. Physicians completing the online tutorial and passing a test at the training platform Curacampus, receive a SDM certificate and credits. Certified doctors are now authorized to invoice for higher counselling rates. Conclusions: A promising and malleable approach to improving SDM in medical consultations has been developed and studied successfully. Whether doktormitSDM can change daily practice is a matter currently undergoing further investigations. Background: Development of skills in research methodology in undergraduate students is a challenge. The Medicine Faculty of Universidad de La Frontera in Chile has implemented integrated and student-centred curricula for all nine programs. These curricula include a line about management and research in health. In the first two years, the students are distributed in multi-professional groups to cover this subject. Objectives: Evaluate the course: ‘Management and Research in Health III’, in the last six years, from the student perspective. Methods: At the end of course, the students evaluate the teaching process through a validated questionnaire. This questionnaire has seven dimensions: organization and responsibility, update and clarity, motivation and student participation, knowledge achieved, interpersonal relationship, evaluations, and comprehensive training. Each dimension includes questions that must be answered on a 1 to 5 scale, where 1 is equal to not achieved, and 5 completely achieved. A descriptive analysis is made for each dimension and for the course. Results: Results from 1764 questionnaires from 2009 until 2014 were included. The total evaluation through years was 4.12. The year with the best score was 2013 with 4.54, and the year with worst evaluation was 2009 with an score of 3.89. The best dimension was comprehensive training with an 4.24 (standard deviation (SD) 0.81), and the dimension with the worst score was knowledge achieved with 4.03 (SD 0.95). A trend of better evaluation through the years was observed. Conclusions: The global students’ evaluation is good, and highlights the value assigned to comprehensive training dimension. These results show that one of the objectives that originated this innovative approach has been fulfilled. P1.023 Effect of evidence-based medicine courses on informatics competencies of medical college students Qin W1 , Chen Y2 , Zhou Q1 , Wang J1 , Chen J1 1 Department of Evidence-Based Medicine and Clinical Epidemiology West China Hospital Sichuan University, China; 2 China National Nuclear Corporation 416 Hospital, Chengdu, China Background: Effective and appropriate use of information and communication technologies is an essential competency for all healthcare professionals. With the development of information globalization, the ability to acquire, process Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Cochrane Database Syst Rev Suppl 1–327 (2015) 173 and use information is particularly important. Objectives: To evaluate the effects of evidence-based medicine (EBM) courses on informatics competencies of medical college students. Methods: We included 88 fifth- and seventh-year medical program and postgraduate medical students who selected the EBM course. We used the self-completion ‘informatics competencies questionnaire’ to describe and measure students’ informatics competencies pre-and post-EBM courses. The questionnaire contains three parts: information acquisition capability, processing capability, and usage capability. Results: Information acquisition capability: students like obtaining information through the Internet, library and communicating with teachers and classmates. After the EBM courses, the students could construct search strategies better, most of the students (56.1%) could formulate answerable clinical questions, and the postgraduate change obviously in expanding database (86.4% to 97.6%, P value < 0.05). Information processing capability: after the EBM courses, the proportion identifying important information increased, and more students could tell the true information from the false (79% to 89.9%, P value < 0.05). Information usage capability: after the course, more students could cite the article by article with right format (27.1% to 45.8%, P value < 0.05). More postgraduates could adopt the true information (35.5% to 48%, P value < 0.05). Conclusions: The evidence-based medicine course has a positive effect on the informatics competencies of medical students. and to provide additional support, if necessary. Results: We hosted three fellowships in 2014, with a total of 16 lead authors and 10 co-authors, representing 11 CRGs. Participant feedback was positive. Factors enabling authors to progress with their reviews included: dedicated time, ongoing technical and methodological support, ability to work in author teams and liaising with CRGs to solve immediate issues. Seven of 16 lead authors submitted reviews for peer review (Table 1). The remaining five authors who had planned to submit by February 2015 have not adhered to their timelines. Reasons include lack of dedicated time when back at work; delays in feedback from co-authors; and underestimating time required to dedicate to the review. Of those that submitted for peer review, additional issues include peer review process delays. Conclusions: Providing fellowships for Cochrane authors to have dedicated time to work on their reviews is an effective strategy for review progress. However, further active support is required post-fellowship as authors continue to experience similar time-management problems. P1.024 Providing time out through dedicated fellowships: strategy to improve Cochrane author support towards review completion Durao S1 , Oliver J2 , Young T3 , Kredo T2 1 South African Cochrane Centre, South Africa; 2 South African Cochrane Centre, South African Medical Research Council, South Africa; 3 Centre for Evidence-based Health Care, Stellenbosch University, South African Cochrane Centre, South African Medical Research Council, South Africa Background: A 2013 survey of 64 Cochrane authors in the South African Cochrane Centre (SACC) reference countries identified limited time and financial constraints as the main barriers to completing reviews. These findings informed a new training strategy to support authors through fellowships for dedicated time to work on their reviews. Objectives: To assess the progress of authors who received fellowships to complete or update their reviews. Methods: Authors in SACC reference countries could apply for fellowships. Eligible authors required a published protocol or a review update due, with identification and selection of studies completed. The five-day programme consisted mainly of time to work on reviews, with short daily methods sessions, ongoing support from Centre and associated staff, and linking with Cochrane Review Groups (CRGs). Participants completed a timeline that the SACC uses for regular follow-up, for project management Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. P1.025 Curriculum reform at Pontificia Universidad Católica de Chile School of Medicine: an opportunity to review the teaching of evidence-based medicine Rivera S1 Pontificia Universidad Católica, Chile 1 Background: In recent years, our institution started a process of curricular analysis, literature review, conducted general needs assessment and targeted needs assessment of learners taking into account: academic advance of students, new laws for health care and patient’s rights, globalization and new trends in medical education. Thus, we realized the curriculum had not evolved enough and graduates might not be prepared to face the future challenges of medical practice. One aspect was to improve the teaching of evidence-based medicine. Objectives: Evaluate our curriculum, make the necessary changes, and adjust the undergraduate profile of our students to the new requirements, including competencies related to decision-making based on evidence. Methods: Kirkpatrick Model for Curricular evaluation: the evaluation process began in 2009, with the formation of various committees to analyze the current curriculum, curriculum relevant international Cochrane Database Syst Rev Suppl 1–327 (2015) 174 Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Cochrane Database Syst Rev Suppl 1–327 (2015) 175 universities, the requirements for physicians in the new health scenarios and projections, pre joints with graduate medical education trends and expert opinion on each of the topics. The views of students, alumni, faculty and the information obtained from the evaluation processes and curriculum accreditation is sought. Results: We designed a new curriculum for a new graduate profile by learning outcomes in eight domains of competencies in a comprehensive framework. One of these skills is related to the management, communication of information and evidence for clinical decision making (see Table 1). This six-year program will be implemented in 2015, and includes improved curricular sequence, content integration, melding basic science and clinical concepts throughout the career, with development of the humanities and professionalism, prioritizing early contact with patients and new methodologies, and early teaching of management of information, evidence and technology in clinical decision making (see Figure 1). Conclusions: Curriculum processes require analysis models to make a proper diagnosis of learning needs, define a undergraduate profile and implement a system evaluation to deliver feedback. P1.026 Informing Cochrane learning strategies: effectiveness of e-learning strategies to increase evidence-based health care (EBHC) competencies Rohwer A1 , Rehfuess E2 , Young T1 1 Centre for Evidence-Based Health Care, Stellenbosch University, South Africa; 2 Institute for Informatics, Biometry and Epidemiology, Ludwig-Maximilians University of Munich, Germany Background: Electronic learning (e-learning) strategies have been widely adopted by educators worldwide. It is a complex intervention with multiple components and dimensions that may interact differently in different contexts. Objectives: To assess the effectiveness of evidence-based healthcare (EBHC) e-learning on EBHC competencies in healthcare professionals. Methods: We considered randomised and non-randomised controlled trials, interrupted time series and controlled before-and-after studies comparing EBHC e-learning or blended learning to no learning, face-to-face, blended or e-learning for healthcare professionals. We searched MEDLINE, EMBASE, CENTRAL, CINHAL, ERIC, SCOPUS, PsychInfo and Web of Knowledge. Two authors independently screened search outputs, selected studies for inclusion, extracted data and made risk of bias judgements. We synthesised results narratively due to large amounts of heterogeneity between studies. Results: After screening 5110 titles and abstracts, we identified 38 potentially eligible studies, 19 of which met our inclusion criteria. Included studies were individually and cluster RCTs and controlled trials. Participants were undergraduate and postgraduate students as well as practicing clinicians. Most studies compared EBHC e-learning to lectures or no learning, some Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. compared blended learning to lectures or no learning. Overall, studies were of moderate quality, but 11 studies had high risk of attrition bias. Knowledge scores improved after e-learning interventions when compared to no learning. Knowledge scores improved in both groups when e-learning was compared to face-to-face interventions, and there was generally no difference in the improvement. One study that compared a clinically integrated e-learning module to a pure online module showed significantly more improvement in knowledge scores with the former. Conclusion: E-learning of EBHC is effective in increasing EBHC knowledge and skills, but should be integrated into clinical settings for maximum benefit and relevance. Cochrane can draw on our findings to inform learning strategies to increase capacity to access, assess and interpret Cochrane Reviews. P1.027 Are Polish physicians familiar with the Cochrane Collaboration and the Cochrane Library? Leśniak W1 , Bala MM1 , Jaeschke RR1 , Koperny M2 1 Jagiellonian University Medical College, Krakow; Polish Branch of the Nordic Cochrane Centre, Poland; 2 Regional Sanitary-Epidemiological Station, Krakow, Poland Background: In October 2014 the Polish Branch of the Nordic Cochrane Centre was established. This entity plans a range of activities, including raising awareness of Cochrane activities and training in the use of Cochrane Reviews (CRs). Objectives: The aim of the study was to assess the awareness of Cochrane and its activities, use of the Cochrane Library (CL) and CRs in Poland and the need for workshops on the use of systematic reviews and literature searching. Methods: A website (using one of the most popular portals for health professionals in Poland) questionnaire-based study. Results: During a four-week period 169 doctors completed the questionnaire: 52% were women; 3% had finished their medical studies less than a year previously, 15% had finished within one to five years, 15% within five to 10 years, 32% within 10 to 20 years, and 34% more than 20 years ago. Seventeen per cent of participants work only as general practitioners, 8% only as specialists in outpatient clinics, 61% in hospitals, and 14% at universities. Most participants (77.5%) had heard about Cochrane, 70% stated that they knew about Cochrane activies, 64% had read publications in CL, 49% had read full texts of CR, 52% used the results of CR in clinical practice (all these results were independent of the place of work or on the time since graduation; there was a statistically significant relationship between the use of CR in clinical practice and reading the full texts of CRs (P value < 0.001)). It is expected that these percentages will be lower in the general population of Polish medical doctors than in the study population, where participants were probably interested in the topic. The frequencies of using CL were: 32% more than once a month, 15% twice a month, 7% once a week, 4% more than once a week, 42% never used it. Eighty-five per cent of Cochrane Database Syst Rev Suppl 1–327 (2015) 176 participants were willing to take part in workshops on the use of CR in clinical practice and searching for clinical information in databases. Conclusions: The awareness of Cochrane activities and products and the use of CL is unsatisfactory in Poland. There is a necessity for promoting use of CR through training activities. P1.029 Increasing utilization of online databases by evidence-based medicine curricula among medical students Chiu Y1 , Kuo K1 , Weng Y2 , Chen C1 , Chen K1 1 Taipei Medical University, Taiwan; 2 Chang Gung Memorial Hospital, Taiwan Background: Evidence-based medicine (EBM) has been identified as a core competence to help physicians improve healthcare quality. One of the important steps to implement EBM is to acquire evidence-based information. Objectives: EBM curriculum was introduced for third-year undergraduate medical students. The current study aimed to determine whether education can enhance their usage of evidence-based online databases. Methods: A structured questionnaire survey was conducted for third-year medical students who received medical curriculum of EBM in 2013. Students completed a baseline questionnaire survey prior to the beginning of curriculum and finished the same questionnaire right after the end of curriculum. Pair-sample t test was used to compare the values of means between the pre- and post-curriculum questionnaires. Results: After the curriculum, students were more likely to access the evidence-based retrieval databases –including two Chinese databases and eight English databases. Specifically, medical students more often accessed the National Digital Library of Theses and Dissertations in Taiwan (P value 0.046), the Chinese Electronic Periodical Service (P value 0.013), the Cumulative Index to Nursing & Allied Health Literature (P value < 0.001), the Cochrane Library (P value < 0.001), MD Consult/ Clinical Key (P value < 0.001), MEDLINE/PubMed (P value < 0.001), ProQuest (P value < 0.001), UpToDate (P value < 0.001), Micromedex (P value 0.011), and DynaMed (P value 0.037). In contrast, the usage of Web portals was not increased (P value 0.380). The common motivations to access the online databases were self-learning and class assignment. Conclusions: The medical curriculum of EBM was helpful for accelerating the utilization of online evidence-based retrieval databases. The data have clinical implications for medical educators to disseminate the implementation of EBM. Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Systematic review methods P1.031 Progress of PROSPERO registering systematic reviews Ciapponi A1 , Glujovsky D1 , Bardach A1 1 Instituto de Efectividad Clı́nica y Sanitaria - IECS, Argentina Background: PROSPERO is an international open access electronic database of prospectively registered systematic reviews in health and social care. A unique registration number and key features from the review protocol are recorded permanently in order to avoid unplanned duplication and enable comparison of reported review methods with what was planned in the protocol. Objectives: To assess the progress of PROSPERO registering systematic reviews, in absolute terms and in relative terms compared to systematic reviews published in MEDLINE. Methods: First, we identified the monthly cumulative registrations and the status of these registrations on PROSPERO (http://www.crd.york.ac.uk/PROSPERO/) from inception date (22 February 2011). Second, we identified the annual publications of systematic reviews in PubMed using the most specific filter developed by Montori et al. (((MEDLINE [tiab]) OR systematic[tiab])) AND ((review[tiab]) OR meta-analysis[Publication Type]) and we also identified PROSPERO registrations (((prospero) AND ((registr* or crd)))). Results: From inception to 28 February 2015 (four years) 6026 registrations were recorded in PROSPERO, most of them (87%) reporting an ongoing status (Table 1). We observed an increasing trend in the number of records (Figure 1). The proportion of systematic reviews published in MEDLINE that reported a CRD record number is also growing but, remains below 2% (Figure 2). Conclusions: PROSPERO registration is growing but is still remarkably low. Hopefully the PRIMA-P 2015 statement will boost this trend further. Cochrane Database Syst Rev Suppl 1–327 (2015) 177 Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Cochrane Database Syst Rev Suppl 1–327 (2015) 178 P1.032 The Embase project 1: crowd characteristics and behaviour after 18 months Noel-Storr A1 , Dooley G2 , Glanville J3 , Foxlee R4 1 Cochrane Dementia and Cognitive Improvement Group, Oxford University, United Kingdom; 2 Metaxis Ltd, United Kingdom; 3 York Health Economics Consortium, United Kingdom; 4 Cochrane Editorial Unit, United Kingdom Background: The Embase project is managed by a consortium made up of Metaxis Ltd, the Cochrane Dementia and Cognitive Improvement Group and the York Health Economics Consortium. The project’s objectives are to identify reports of randomised trials in Embase and to submit those reports to Cochrane’s Central Register of Controlled Trials (CENTRAL). The project has been managed by this consortium since April 2013 and uses crowdsourcing for much of the screening. Objectives: This poster will focus on presenting answers to the following three questions about the screening crowd: 1. What do we know about the crowd? 2. Do screeners become more confident over time? 3. How engaged is the crowd? Methods: By September 2015 the project will have generated over 18 months’ worth of data. Previous publications have focussed on crowd performance, but here we focus on the characteristics and behaviour of the crowd. We will analyse the data from the project’s sign up form, which all screeners must complete. This includes demographic information plus information about prior knowledge of randomised trial design. We will assess screeners’ decision-making over time by looking at: time per citation and the proportion of ‘Unsure’ classifications made. To determine screeners’ engagement we will look at screeners’ average activity per month/year. Results: The 18-month results will be presented in this poster. Interim data suggest that the crowd is 57% female; that most of the crowd is aged between 31 and 40 years old; and around 4% were not familiar with randomised trial design before taking part. Conclusions: Crowdsourcing brings with it many potential benefits, but is not a simple option. The more we can learn about our participants, the more we can help make the experience rewarding and fulfilling for them and retain their involvement in this project and others that are now underway, such as Project Transform. P1.033 Experiences with Covidence in preparing a comprehensive systematic review Helfer B1 , Samara M1 , Leucht S1 1 Cochrane Schizophrenia Group, Germany Background: Covidence is ‘a not-for-profit service dedicated to improving the use of evidence in health-care decision making’ (www.covidence.org). It is an on-line platform to perform systematic reviews and meta-analyses. Objectives: To access the usefulness of Covidence in performing a comprehensive systematic review and meta-analysis. Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Methods: In 2014 to 2015 we performed a systematic review and meta-analysis of antidepressants added to antipsychotic drugs for schizophrenia using Covidence. This included comprehensive screening (approximately 6500 abstracts), full text review (approximately 250 full text articles), data extraction and assessment of risk of bias (approximately 100 included studies) and export of data into Review Manager software (RevMan). All steps were performed by two independent reviewers. Based on our own experiences with the software and correspondence with the Covidence support team, we analyzed the advantages and disadvantages of this service. Results: Although initially plagued with numerous technical issues and missing functionalities, thanks to continuous improvement based on user feedback, Covidence became an invaluable tool in preparing our systematic review and meta-analysis. As of April 2015, some problems of a technical nature still exist, especially when handling big data-sets, but they are typically easy to overcome with the quick help from the Covidence support team. An important advantage of the software is that all data are safely stored and automatically updated as entered and that Covidence enables reviewers to work independently at the same time and in different geographical locations. Moreover, Covidence is currently free of charge and supports exporting data into RevMan. Conclusions: Covidence can be a very useful tool for systematic reviewers and meta-analysts and their teams. As it is constantly improved, it holds promise for the future. P1.035 Updating systematic reviews published in Chinese: a cross-sectional study Wei D1 , Zhang H2 , Li L3 , Li R2 , Chang X4 , Fang Z3 , Chen Y1 , Yang K1 1 School of Basic Medical Sciences, Lanzhou University; Key Laboratory of Evidence-Based Medicine and Knowledge Translation of Gansu Province, Lanzhou, China; 2 The Second Clinical Medical College of Lanzhou University, Lanzhou, China; 3 School of Basic Medical Sciences, Lanzhou University, Lanzhou, China; 4 The First Clinical Medical College of Lanzhou University, Lanzhou, China Background: Updating systematic reviews (SR) is a useful solution for reducing the impact of publication bias (or time lag bias) on the results. Nevertheless, the authors of SRs do not keep SRs up to date. Research shows that only 3% of SRs published in peer-reviewed journals and 38% of Cochrane Reviews have been updated. The quantity of SRs published in Chinese is increasing every year. However, there is little study that focuses on whether they are updated. Objectives: To investigate the current situation regarding the updating of systematic reviews published in Chinese. Methods: We used the terms ‘systematic review’, ‘meta-analysis’, and ‘update’ to search two electronic databases, WANFANG Data and China National Knowledge Infrastructure (CNKI). Meanwhile, we performed a handsearch of the four Chinese journals with ‘evidence-based’ in their titles. The previous version of SRs were also identified. Two independent reviewers Cochrane Database Syst Rev Suppl 1–327 (2015) 179 screened the updating of SRs and extracted the data. The data extracted included publication year and journal, search strategy, etc. Results: Five updated SRs were included. Of those, three (60%) were published in peer-reviewed journals, and the two (40%) remaining SRs were published as an abstract of conference and an academic dissertation. Only one (20%) SR reported they had updated their own SR. The range of updating period was from one to seven years. The conclusion had changed in one (20%) SR. Conclusions: A few SRs published in Chinese have been updated; the period between initial publication and updating varies and is generally long. In addition, generally the authors of SRs do not consider updating their own SRs. P1.036 It’s the noise that makes the poison: why current methods for assessing chemicals for toxicity provides so little evidence? Mandrioli D1 , Silbergeld E1 1 Johns Hopkins Bloomberg School of Public Health, USA Background: Toxicology as a field has not often produced efficient and timely evidence for decision making in public health. In fact, of the 87,000 chemicals registered for commerce in the USA, only one-tenth have been tested for potential harms. Several in vitro and in vivo tests have been adopted, without appropriate validation, and the process of hazard assessment is extremely slow and largely based on non-transparent practices such as ‘expert opinion’ and ‘weight of evidence’. In response to this, the US National Research Council called for the adoption of evidence-based methods and systematic reviews in regulatory decision making. EPA, FDA and the European Food Safety Agency have recently endorsed these methods in their assessments of safety and risk. Objectives: Evaluating the effects of the adoption of evidence-based methods in toxicology for filtering the primary literature by developing criteria to identify tests of low quality and high risk of bias and thereby accelerating hazard assessment. Methods: We compare current practices in toxicology with principles and methods utilized in evidence-based medicine and health care, with emphasis on Cochrane’s record. Several aspects are analyzed: validity and quality of the evidence, transparency, time for evaluation, unnecessary animal loss, consistency of the evaluations. Results: Evidence-based toxicology (EBT) methods could provide an effective filter of low-quality studies and reduce the overall time and resources needed for risk assessment in the future. EBT methods proposed by OHAT and Navigation Guide represent the first step in this direction, although further validation of the methods is necessary. Adherence to Cochrane principles, as well Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. as transparent development of systematic methods, is a fundamental prerequisite for future implementation of EBT. Conclusions: The adoption of consistent principles and methods is likely to enhance the validity, transparency, efficiency and acceptance of toxicological evidence, with benefits in terms of reducing the burden of animal used, delays and costs for all stakeholders (researchers, consumers, industry). P1.039 Development of an algorithm for choosing study designs Peinemann F1 , Kleijnen J1 1 Maastricht University, The Netherlands Background: The purpose of systematic reviews is to answer distinct research questions on healthcare interventions. The type of study design, which is scheduled for inclusion in a systematic review, may be critical for its findings and consequences. We did not find an existing algorithm providing guidance for systematic reviewers to choose the most appropriate study designs. Objectives: The aim of this study was to develop an algorithm that provides guidance and awareness for various design characteristics suitable to answer distinct research questions in healthcare interventions. Methods: We addressed the following questions. What is the preferred content and format of a straightforward algorithm for integration of various study designs in a systematic review that incorporates major study characteristics as decision points and that allows consideration of practical concerns? Can the algorithm hold its promise as a useful guide for choosing appropriate study designs by testing its feasibility and applying it to existing systematic reviews? Results: We developed an algorithm that incorporates length of follow-up (long, short) and frequency of events (rare, frequent) as binary decision points. At the end of these four pathways, each of five types of outcome (death, disease, discomfort, disability, dissatisfaction) was linked with at least one design label. The algorithm also considers practical or ethical concerns as well as unavailable best evidence, and it provides examples for explaining the critical guidance points. We confirmed the usefulness of the pathways by reproducing the study selection and its impact on the conduct of four systematic reviews. Conclusions: The algorithm facilitates decisions about inclusion of different study types to those planning to prepare a systematic review. The algorithm cannot be applied without the consideration of disease-specific circumstances and aims of interventions. However, we think that it may reduce the time burden on review authors and may increase the scope and applicability of systematic reviews. Cochrane Database Syst Rev Suppl 1–327 (2015) 180 P1.040 Importance of author inquiries Peinemann F1 1 University of Cologne, Germany Background: Systematic reviews may rely on study data published in peer-reviewed journals. A good international reputation and a high impact factor of a journal may promise that the article contains correct and verified data. We conducted a systematic review on drug intervention in patients with high-risk neuroblastoma and included one study for which we found some inconsistencies within the main report published by the Journal of Clinical Oncology (impact > 17). Methods: I found that it was difficult or not possible to reproduce the results of the statistical analyses of the included study. I did not know if I just needed help to understand the data better, whether I faced minor typing errors, or if I had discovered flaws serious enough to change the conclusion of the study. Therefore, I located the authors and asked them to reply to my inquiries by email. Results: I am very grateful to the authors for leaving room for my doubts. The authors took great efforts to re-evaluate the analyses thoroughly. Finally, the resulting changes were published as an erratum, which was attached to the online version of the article. The difference of overall survival between the treatment groups was not significant as opposed to the earlier version. Conclusions: The inquiry of authors to explain their analyses is important and can have an impact on the study conclusion and possibly the treatment of patients. Respectful appreciation of the work achieved by the authors is a prerequisite. Highly skilled professional scientists may reply appropriately and pave the way effectively for clearing the data. P1.041 The importance of registration for systematic reviews/meta-analysis about traditional Chinese medicine in Chinese 1 1 1 1 2 3 3 4 Wang J , Zhang L , Mu W , Liu Z , Hu J , Du L , Li Y , Shang H 1 Ianjin Branch of the Ministry of Education Virtual Research Center for Evidence-Based Medicine, China; 2 Chinese Medicine Institution of Beijing, Beijing Hospital of TCM, China; 3 Chinese Cochrane Center, China; 4 Key Lab of Chinese Internal Medicine of the Ministry of Education, Dongzhimen Hospital, Beijing University of Chinese Medicine, China Background: With the rapid development of evidence-based medicine in the field of traditional Chinese medicine (TCM), a lot of low quality systematic reviews have been produced, which may burden the evidence and affect decision-making. Objectives: In order to explain the importance of registration, this research studies the current situation regarding it. Methods: Systematic searching of CNKI, VIP, WanFangdatabase, CBM, PubMed, Web of Science (WOS), the Cochrane Library and the PROSPERO registry platform was Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. carried out to included all published SR/MAs about TCM in both Chinese and English. General information was extracted individualy and the report and methodological qualities were assessed independently by PRISMA and AMSTAR checklists. According to data from CNKI and WOS, cited information was identified. Results: A total of 2460 SR/MAs have been published since 1997 and only 10.85% were published before 2008. All 56 registered research with full text came from the Cochrane Library. The number of SR/MAs written by Chinese people was about 8.36 times (1958/234) more than non-Chinese. However, the number of high methodological and report quality SR/MAs in Chinese are less than 1/4 (19/80) in English. Although the Chinese SR/MAs involved 179 kinds of diseases, 28.72 %"i566/1971) concentrated on five single diseases while only 18.20 %"i89/1971) in English which brought high recurrence rate of same subject in Chinese. 70.98% Chinese SR/MAs were cited for 7990 times in total, accounting for 5.71 times on average. While the ratio was 67.69% (331/489) in English with total 4903 citations and 9.98 times on average. Within these 331 cited English articles, 289 of them have been cited by Chinese writers in a total of 778 times, which equalized to 2.69 times per articles. Conclusions: Although the number of TCM SRs increases rapidly, low quality, disease-concentration, outcomes bias of the studies cannot be ignored. Registration may helpful to control the overload evidence and the misleading, especially important for articles in Chinese, because TCM clinicians prefer to produce and use the Chinese SR/MAs due to the language preference and evidence accessibility. P1.042 Does applicability make the difference? Two systematic reviews on the same topic Angelescu K1 , Nussbaumer B2 , Sieben W1 , Scheibler F1 , Sauerland S1 1 Institute for Quality and Efficiency in Health Care (IQWiG), Germany; 2 Department for Evidence-Based Medicine and Clinical Epidemiology, Danube University Krems, Austria Background: Most guidelines recommend screening for and treatment of asymptomatic bacteriuria (ASB) in pregnancy. A Cochrane Review (CR) on antibiotic treatment of ASB in pregnancy provides the basis of this recommendation. However, an Institute for Quality and Efficiency in Health Care (IQWiG) report on benefits and harms of screening for ASB concluded that the benefit is not proven. Objective: We compared the two systematic reviews to find out why they have come to differing conclusions. Methods: We compared PICO, study pool, risk of bias (RoB) assessment, effect estimates, assessment of applicability, and conclusions. Results: Differences in inclusion and exclusion criteria result in different study pools. The IQWiG report included only 3 of 14 studies that were included in the CR. However, effect estimates of pyelonephritis rates are roughly comparable (CR: relative risk = 0.23, 95% confidence interval (CI) 0.13 to 0.41; IQWiG: odds ratio = 0.21, 95% CI 0.07 to 0.59). Studies date back from the 1960s to the 1980s. Both reviews rate Cochrane Database Syst Rev Suppl 1–327 (2015) 181 the RoB of included studies as high. Both do not describe a methodology to assess applicability. However, in the CR applicability is discussed with respect to study medication and microbiologic methods, but not generally rejected. The IQWiG report regards interventions in 2 of 3 included studies as not applicable to current health care. In conclusion, the CR identifies preventive effects, but recommends interpreting the results cautiously. The IQWiG report, which, unlike the CR, comprises a dichotomised statement on benefit and harms, regards the benefit of an ASB treatment as not proven due to serious concerns regarding applicability. Conclusion: Despite differences in detail, overall conclusions regarding the effect of treatment do not differ much. Both works see the need of further randomised trials. The main difference is seen in the significance of applicability and derived conclusions. To date there is no well-developed methodology for assessing applicability. However, different domains of applicability (population, setting, interventions, outcomes, follow-up) should be assessed as recommended in current EUnetHTA guidance. P1.043 Exploring inconsistencies between observational and experimental studies of selenium and diabetes risk Vinceti M1 , Filippini T1 , Del Giovane C1 , Crespi C2 1 University of Modena and Reggio Emilia, Italy; 2 Fielding School of Public Health, UCLA, USA Background: Observational and experimental epidemiologic studies that have addressed the relation between intake of the trace element selenium and cancer risk have yielded strongly conflicting results, as recently reported by a Cochrane Review. Most observational studies suggest an inverse association, while randomized controlled trials (RCTs) have indicated a null or direct relation. Little is known about the replication of such inconsistencies when dealing with the risk of other chronic disease. Objectives: We investigated the results of observational and experimental studies linking selenium exposure to the occurrence of type-2 diabetes. Methods: After a literature search, we identified 12 observational studies (eight cross-sectional and four cohort) and five RCTs. Using a random-effects model, we computed the summary relative risk (RR) of type-2 diabetes along with its 95% confidence interval (CI) in subjects with the highest versus the lowest selenium exposure category in observational studies, and in subjects allocated to selenium compared to placebo in the RCTs. Results: Summary RRs were 1.98 (95% CI 1.22 to 3.23) and 1.13 (95% CI 0.15 to 8.45) for cross-sectional studies using serum and toenail selenium for exposure assessment, respectively. Cohort studies based on toenail selenium yielded a summary RR of 0.68 (0.72 to 0.98), while the only study assessing dietary selenium intake gave a RR of 2.39 (1.32 to 4.32). For RCTs, summary RR was 1.10 (1.00 to 1.21) among selenium-supplemented versus placebo. The distinctive feature of the two observational studies (one cross-sectional and one prospective) that failed to find an excess diabetes risk associated with higher selenium exposure was that the Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. subjects were health professionals. Age, gender, study area and other demographic characteristics did not appear to have influenced the results. Conclusions: These results suggest that the ability of observational studies to predict results of RCTs when addressing the health effects of selenium may differ on the basis of the outcome studied (diabetes versus cancer) as well as the indicator used for exposure assessment and the type of population under study. P1.045 Project on a Framework for Rating Evidence in Public Health (PRECEPT): structure of a draft framework Harder T1 , Takla A1 , de Carvalho Gomes H2 , Eckmanns T1 , Ellis S3 , Forland F4 , James R5 , Jansen A2 , Meerpohl J6 , Morgan A3 , Rehfuess E7 , Schuenemann H8 , Wichmann O1 , Zuiderent-Jerak T9 1 Robert Koch Institute, Germany; 2 European Centre for Disease Prevention and Control (ECDC), Sweden; 3 National Institute for Health and Care Excellence (NICE), United Kingdom; 4 Norwegian Institute of Public Health, Norway; 5 Scottish Intercollegiate Guidelines Network (SIGN), United Kingdom; 6 German Cochrane Center, Germany; 7 University of Munich, Germany; 8 McMaster University, Canada; 9 Linköping University, Sweden Aims: The Project on a Framework for Rating Evidence in Public Health (PRECEPT) is a collaboration between European public health agencies and universities, established in 2012, that aims to establish a framework for evaluating and grading evidence in the field of infectious disease epidemiology, prevention and control. PRECEPT is funded by the European Centre for Disease Prevention and Control (ECDC). This presentation describes the structure and workflow of a draft framework. Methods and Results: The PRECEPT framework is designed to rate scientific evidence related to four domains that are of high priority in infectious disease prevention and control: disease incidence/prevalence, risk factors for disease, diagnostics and intervention. The framework is grouped into six consecutive working steps, starting from a complex public health question and ending with an evidence statement for each relevant domain. In step 1, approaches are described for identification of relevant questions. In step 2, methodological guidance is provided for the conduct of systematic reviews for these questions. For the appraisal of methodological quality of identified individual studies, 15 different quality appraisal tools are proposed and an algorithm is given to match a given study design with an appropriate tool (step 3). In step 4, a generalized evidence grading scheme based on the GRADE methodology is provided to rate the quality of bodies of evidence for each domain. The evidence appraisal process ends with the preparation of evidence profiles and summary of finding tables (step 5) followed by preparation of an evidence summary for communication of the results (step 6). By applying this methodological framework, the user should be able to evaluate and grade scientific evidence from the four major domains in a transparent and reproducible way. Outlook: The draft framework is currently being piloted Cochrane Database Syst Rev Suppl 1–327 (2015) 182 by applying it to examples from infectious disease prevention and control. Further refinements of the methodology, as well as application to other domains, and the inclusion of a methodology for going from evidence to recommendations are planned for future versions of the framework. P1.046 The unique challenges of systematic reviews and meta-analysis of cluster-randomized trials Chen Y1 , Yao L1 , Wang X1 , Wei D1 , Wang Q1 , Wang M1 , Yang K1 1 Lanzhou University/Chinese GRADE Center, China Background: Cluster-randomized controlled trials (CRTs) have greatly increased over the past 20 years and have motivated the publication of an extension of the CONSORT statement for this design because of its particular methodological issues. However, there are few systematic reviews (SRs) or meta-analysis of CRTs that have been conducted or published. Objectives: To address key questions when conducting SRs of CRTs. Methods: We searched MEDLINE, Embase, the Cochrane Library, CNKI (China National Knowledge Infrastructure/Chinese Academic Journals full text Database), VIP (a Chinese full-text database), WANFANG (a Chinese full-text database) and CBM (China Biomedicine Database Disc) using the term cluster*, group*, field*, community, communities, systematic review*, meta analysis in titles on 16 January 2015. We also checked the section about CRTs in the Cochrane Handbook for Systematic Reviews of Interventions. Results: We identified 696 papers but only four were relevant to SRs of CRTs. The main challenges of systematic reviews and meta-analysis of CRTs are how to: 1. develop highly sensitive search strategies for SRs of CRTs; 2. assess the particular risk of bias of CRTs; 3. deal with CRTs with different types of clusters; 4. choose statistical methods to combine the results; 5. combine the findings from CRTs; 6. conduct the subgroup analysis; 7. consider intracluster (or intraclass) correlation coefficient (ICC); 8. GRADE the evidence, such as recruitment bias, in CRTs; 9. explain the difference when cluster- and individually randomized trials in an SR are inconsistent for the same topic; 10. generalize the findings. Conclusions: The SOC(systematic reviews of cluster randomized trial) Working Group has already explored highly sensitive search strategies for SRs of CRTs. We expected to work together with methodologists, researchers using CRTs and statisticians to develop a more detailed guideline for systematic reviewers of CRTs. Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Quality of reporting P1.047 A checklist approach to evaluation of scientific study protocols approval removes bottleneck and improves methodological quality: a randomized controlled trial Kumar A1 , Kosambiya J2 , Reljic T1 , Chenneville T1 , Menezes L1 , Mhaskar R1 , Miladinovic B1 , Djulbegovic B1 1 Morsani College of Medicine, USF Health, Program for CER, Tampa, USA; 2 Government Medical College, Surat, India Background: Scientific Review Committee (SRC) members mostly use qualitative methods for reaching decisions (approval or disapproval) on study protocols. Checklists for reporting scientific studies have helped improve transparency and presentation of studies in the public reporting of research findings. While the use of checklists has become prevalent in the arena of publishing, their role in the SRC evaluation and decision-making process has not been assessed. Objective: To assess the efficacy of the SRC approval process using an objective checklist in addition to the current procedures versus standard operating procedures (SOP) alone for making a decision on approval or disapproval of scientific study protocols. Methods: Two committee members were randomly assigned to review each protocol involving human subjects using the checklist plus the SOP or SOP alone. Checklists (e.g. CONSORT, STROBE etc.) recommended by scientific journals were matched to study design. The primary outcome was approval rate for checklist(s) versus no checklist(s) group. Differences in disapproval rates between the two groups were assessed using a Chi2 test. The agreement between reviewers (checklist versus no checklist) was assessed using the Kappa statistic. Results: A total of 28 study protocols were reviewed in duplicate by 56 reviewers (10 participants were excluded due to incorrect review). Characteristics of reviewers in compared groups were similar (see Table 1). The relative risk reduction in disapproval was 33% with use of checklist plus SOP compared with SOP alone (absolute risk reduction = 25%; relative risk = 0.67; 95% CI 0.44 to 1.03). The agreement between two reviewers (checklist versus no checklist) assessing the same protocol was poor, resulting in kappa of 0.07 (95% CI 0 to 0.39). Conclusion: Use of checklist along with SOP results in lower disapproval compared with the SOP only. The lack of statistical significance is likely due to small sample size. Introduction of checklists in the study design phase ensures better methodological quality of study protocols and assists SRC members remove the bottleneck in approvals and serve as facilitators of research. Cochrane Database Syst Rev Suppl 1–327 (2015) 183 review and meta-analysis, resulting in wasted information. Objectives: To quantify the proportion of incomplete outcome data reporting for randomized controlled trials (RCTs) of medical interventions for open-angle glaucoma (OAG). Methods: We searched electronic databases for RCTs of medical interventions for OAG with no restrictions with respect to date or language. Two individuals independently screened records and abstracted data. For this project, we analyzed the completeness of IOP and visual field (VF) data in included trials. Discrepancies were resolved by consensus. We used the most complete reporting of outcomes for each RCT. Results: We identified 417 trials with 65,452 participants as of May 2014, of which 406 (97%) trials and 87 (21%) trials reported IOP and VF as an outcome, respectively (Table 1). Data from about 56,000 participants from 356 trials was ‘wasted’ due to lack of VF outcome data (Table 1). For IOP, a surrogate outcome, data from about 10,000 participants from 67 trials was ‘wasted’ due to lack of IOP outcome data. Conclusions: Only a small proportion of all identified trials reported complete VF data for meta-analysis, a patient-centered outcome. Investigators have a moral obligation to their participants and colleagues to be more vigilant about complete reporting outcomes and not ‘waste’ data. The comparative effectiveness of glaucoma interventions could be better determined with more complete reporting on VF and other patient-centered outcomes. P1.049 P1.048 Over 56,000 participants’ data ‘wasted’: an example from randomized controlled trials of medical interventions for open-angle glaucoma Law A1 , Lindsley K1 , Rouse B1 , Wormald R2 , Dickersin K1 , Li T1 1 Cochrane Eyes and Vision Group US Satellite, USA; 2 Cochrane Eyes and Vision Group Editorial Base, United Kingdom Background: A completely specified outcome requires the following five elements: domain (e.g. intraocular pressure (IOP)), specific measurement (e.g. applanation tonometry), specific metric (e.g. change from baseline), method of aggregation (e.g. mean), and time-point (e.g. three months). To perform a meta-analysis of the effect of an intervention, one also needs to know the estimate of treatment effect as well as its precision (e.g. standard error) from each trial. Incomplete reporting of an outcome prevents its full use in a systematic Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Quality of published randomized controlled trials (RCTs) in Thai healthcare journals Pengput A1 , Pattanittum P2 1 Master Degree in Public Health Student, Biostatistics Program, Faculty of Public Health, Khon Kaen University, Thailand; 2 Department of Biostatistics and Demography, Faculty of Public Health, Khon Kaen University, Thailand Background: A well-designed randomized controlled trial (RCT) provides the most reliable evidence of the effectiveness of interventions –the gold standard for clinical trials. Many RCTs are published each year with either trivial or serious flaws. Some serious flaws lead to misleading conclusions or failure to provide important information to guide clinical decision making. The more flaws, the lower the quality of RCTs. A way to assess the quality of RCTs is to consider the reporting quality; the CONSORT (Consolidated Standards of Reporting Trials) statement is a tool. There is no research that evaluates the reporting quality of RCTs published in Thai healthcare Journals. Objectives: To describe the quality Cochrane Database Syst Rev Suppl 1–327 (2015) 184 of published RCTs in Thai healthcare journals. Methods: We searched for Thai-language RCTs published in Thai healthcare Journals from January 2008 to December 2012 by using an Online Public Access Catalog (OPAC) of Khon Kaen University and Thai University Library Network of OPAC. For each full RCT text we assessed the reporting quality by means of a questionnaire based on the revised version of the CONSORT statement 2010 (a 25-main item checklist; consists of 37 items). We recorded the quality of reporting presented by each item of the CONSORT checklist, and also the overall quality of included RCTs. Results: The search identified 757 RCTs, 35 of which met the inclusion criteria. Result by each item of CONSORT: 18 of 37 (48.6%) items were reported in at least 75%. Some important items were inadequately reported; item 7a sample size determination (40%), item 8a sequence generation (57%), item 9a allocation concealment mechanism (43%), item 11a blinding (57%), item 13a participant flow (9%), items 17a and 17b reporting of estimated effect size and its precision (14%, 13%). The least reported items were item 23 registration number and name of trial registry (1/35), item 24 where the full trial protocol can be accessed (1/35), item 12b statistical methods for additional analyses (2/14). The overall quality was moderate; 28 RCTs reported on 19 to 27 items. Conclusions: The quality of reporting RCTs in Thai healthcare journals needs to be improved. P1.050 and sentences; - reporting of setting, recipient, provider, procedure, materials and intensity; - use of descriptive aids such as tables and access to supplementary documents. We compared the trials’ description of their intervention between the groups according to whether the intervention was found to be effective, and whether it was published in a leading general medical journal. Results: We have extracted intervention content of the trials. Overall reporting of the content was variable. Early results suggest that trials published after 2008 generally provide longer descriptions of the intervention of interest than those published before 2008 (median word length pre 2008 was 236, post 2008 it was 296). Conclusions: To our knowledge, this is the first comparative study of standards of historical and current intervention reporting. Our results will identify areas for improvement. The result may inform future guidelines. Reference Nieuwlaat R, Wilczynski N, Navarro T, Hobson N, Jeffery R, Keepanasseril A, Agoritsas T, Mistry N, Iorio A, Jack S, Sivaramalingam B, Iserman E, Mustafa RA, Jedraszewski D, Cotoi C, Haynes RB. Interventions for enhancing medication adherence. Cochrane Database of Systematic Reviews 2014, Issue 11. Art. No.: CD000011. DOI: 10.1002/14651858. CD000011.pub4. P1.051 Reporting of interventions: a comparative case study before and after the publication of the extended CONSORT guidelines for reporting non-pharmacological trials 1 1 1 1 Candy B , Vickerstaff V , Jones L , King M 1 University College London (UCL), United Kingdom Background: Limited description of interventions in trial publications wastes resources and potentially harms patients. Interventions cannot be replicated or assessed for generalizability if they are poorly described. Even if found to be ineffective, complete description of an intervention is needed to reduce resource waste and inform approaches. Recognition of the need to improve reporting is not new. Both the 2008 extension to CONSORT statements for non-pharmacological interventions, and the 2010 revised CONSORT statement emphasized the importance of full reporting of the details of an intervention. Objectives: We explore whether intervention reporting is improving. We compare descriptions of similar non-pharmacological interventions in trials undertaken before and after 2008. Methods: We reviewed 182 trials included in a Cochrane Review on interventions for enhancing therapy adherence (Nieuwlaat 2014). We included for increased similarity only psychotherapeutic and educational multi-component interventions. We split the trials into two groups: (1) those published before 2008; and (2) those published after 2008. Two authors independently assessed intervention description. We compared description according to: - number of words Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Misleading reporting in major fertility journals: poor use of confidence intervals and absolute differences Glujovsky D1 , Sueldo C2 , Borghi C2 , Nicotra P2 , Andreucci S2 , Ciapponi A1 1 IECS - Instituto de Efectividad Clı́ nica y Sanitaria, Argentina; 2 CEGYR - Centro de Estudios en Genética y Reproducción, Argentina Background: To get a correct interpretation of the findings in a study, the authors should use correct measures to report them. Although there are lots of items recommended by reporting guidelines, some, specifically, could have a high-impact on the key message for the reader. Some of them (i.e. P value and relative risks) are more frequently used than others (confidence intervals (CI) and interpretation of absolute risks differences), and missing these others provides an incomplete overview of the clinical scenario. Objectives: To evaluate the proportion of randomized controlled trials (RCTs) published in the top fertility journals that reported their results providing the estimated effect size and its precision; and for binary outcomes, the presentation of both absolute and relative effect sizes. Additionally, we did a descriptive analysis of the interpretation of those results. Methods: All the RCTs published in top fertility journals in 2014 were analyzed by pairs of independent reviewers, evaluating the titles and abstracts of identified articles. As a second step, two randomly selected independent reviewers, assessed each included study to finally include them in the analysis and to do the data extraction. Discrepancies were resolved by consensus. Cochrane Database Syst Rev Suppl 1–327 (2015) 185 We analyzed the abstract and the results section of the full text separately to see if the authors mentioned the P value and a CI for the main outcome and for secondary outcomes. For binary outcomes, we evaluated if any relative measure (i.e. RR, odds ratio or relative risk reduction) was used and if any absolute measure (i.e. absolute risk reduction or number needed to treat) was mentioned. For continuous measures, we evaluated if the mean difference and its CI was used. We analyzed whether any reference related to any clinically relevant result was mentioned in the Materials and methods section. We also evaluated the Discussion and Conclusions and described the authors’ interpretation from the results that they published. Results: While more than 90% reported P value, less than 50% reported CI and absolute measures. More results and conclusions will be shown at the Colloquium. P1.052 Critical reporting elements for publications of research and systematic reviews of complex interventions Guise J1 , Viswanathan M2 , Chang C3 , Butler M4 , Kondo K5 , Dickinson C1 , Arbues F5 , Motu’apuaka M5 , Pigott T6 , Tugwell P7 1 US Cochrane West, Oregon Health, Science University, USA; 2 Research Triangle Institute, UNC RTI Evidence-Based Practice Center, USA; 3 Agency for Healthcare Research & Quality, USA; 4 University of Minnesota Evidence-Based Practice Center, USA; 5 Scientific Resource Center for the AHRQ Effective Health Care Program, USA; 6 Loyola University Chicago, USA; 7 Cochrane Health Equity Field/Campbell Equity Methods Group, Canada Background: Complex interventions are widely used in healthcare and public health and are increasingly the subject of systematic reviews. Inconsistencies and shortcomings in the reporting of complex interventions hamper systematic review efforts. Objectives: To develop a standard list of critical elements to report in publications of research and reviews of complex interventions. Methods: We reviewed the literature, interviewed international experts, and conducted a modified Delphi consensus process among international experts in evidence reviews, research, and implementation of complex interventions to generate a final checklist of reporting items for publications of complex interventions. Results: Items considered for reporting of research and reviews of complex interventions included: intervention characteristics such as the number and type of components, adaptation over time, active and optional components, independence versus dependence among components, theoretical, conceptual, and/or functional foundation for the intervention, governance, skills and training of agents delivering the intervention; population characteristics including the intended recipient (e.g. individual, population, health system), cultural, educational, demographic, medical, and skills required to participate; setting/context characteristics including organizational readiness to change or adoption of intervention, key attributes of champions; comparator characteristics including describing ‘usual care’ and rival Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. activities; and outcomes including expected outcomes, unintended consequences, resource utilization, and cost. We will report Delphi results including the final reporting elements checklist. Conclusions: Space is limited in publications yet sufficient details are required to enable wide-scale successful adoption. This checklist may help authors with consistent language and reporting of individual studies, help systematic reviewers focus on key considerations, and inform future research. P1.053 Developing a Preferred Reporting Items for Development of Evidence based Clinical Practice Guideline in Traditional Medicine (PRIDE-CPG-TM) Choi J1 , Jun JH1 , Lee JA1 , Choi TY2 , Lee MS2 , Shin B3 , Choi J2 , Choi J2 1 Korea Institute of Oriental Medicine, South Korea; 2 Korea Institute of Oriental Medicine, South Korea; 3 School of Korean Medicine, Pusan National University, South Korea Background: The development of evidence based clinical practice guidelines (CPGs) is a challenge for all guideline developers in traditional medicine. Using a publication platform with reporting items may overcome these challenges. Objectives: The aim of this study was to develop a unified methods platform with preferred reporting items for development of evidence-based (PRIDE) clinical practice guidelines (CPG) in traditional medicine (TM). Methods: We searched the literature and guideline handbooks, and collected information from methodological conferences. A group of experts edited drafts of the PRIDE, identifying for each item of one or more exemplars of good reporting, and developed text explaining the rationale and discussing relevant evidence. To develop the PRIDE, we also used general and various guideline checklists and drew on the experience gained from Oriental medical doctors and methodological experts. Results: We designed the Preferred Reporting Items for Development of Evidence-based Clinical Practice Guidelines in Traditional Medicine (PRIDE-CPG-TM), in the form of a checklist with descriptions of items and examples, to improve reporting of CPG in TM, and so facilitate their interpretation and replication. The PRIDE included five items and 40 sub-items. These items pertain to the development methodology (22 items), overview of diseases and symptoms (six items), recommendations (four items), implementation and dissemination (five items) and the appendix (three items, i.e. glossary, references etc.). Conclusions: The completeness of descriptions in CPGs is very poor. Therefore, a complete description of the recommendation of TM in CPGs is necessary for physicians to be able to use the recommendations in areas of clinical practice. The PRIDE will be useful guidance for TM developers when developing evidence-based guidelines. Cochrane Database Syst Rev Suppl 1–327 (2015) 186 Conclusion: Substantial room for improvement remains for the reporting of surgical RCTs. A guideline that address issues specific to surgical RCTs should be developed by methodologists, journal editors and surgeons. P1.056 P1.055 Huang T1 , Tam K2 Departmant of Nursing, College of Medicine and Nursing, HungKuang University, Taiwan; 2 Department of Surgery, Taipei Medical University, Shuang Ho Hospital, New Taipei, Taiwan Comparison of the attention paid to reporting guidelines and trial registration in nursing journals in the last three years 1 How many reports you can trust? A cross-sectional study to characterize the quality of research reporting 1 1 1 2 1 Zhang L , Zhao P , Jia P , Chen Q , Zhang M 1 Chinese Evidence-Based Medicine Centre, West China Hospital, Sichuan University, China; 2 West China School of Public Health, Sichuan University, China Background: Muir Gray once said ‘In the 19th century health was transformed by clear, clean water. In the 21st century health will be transformed by clean, clear knowledge.’ Nowadays, substantial articles are published in various magazines. Unless research is adequately reported, the time and resources invested in its conduct is wasted. Objectives: To use the reporting guidelines Consolidated Standards of Reporting Trials (CONSORT) and Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) to compare the quality of reports between 2004 and 2014. Methods: We conducted a systematic review (SR) of 400 randomized controlled human trials and 100 SRs published in MEDLINE, Embase and the Cochrane Library. We randomly selected eligible pieces of research until their number met our required sample size. According to reporting guidelines CONSORT and PRISMA, we assessed these randomized controlled trials and SRs to one of three grades (completely compliant with requirements; generally compliant with requirements, and non-conformant with compliance requirements). Teams of two reviewers will independently screen full texts of reports for eligibility, and abstract data, using standardized, pilot-tested extraction forms. We will conduct univariable and multivariable logistic regression analyses to examine the association of pre-specified study characteristics (such as funding sources, clinical area, type of intervention, and so on) with the quality of research reporting. Progress and Discussion: Currently, we are screening these trial and SR reports. Reporting guidelines such as CONSORT, PRISMA aim to improve the quality of research reports, but all are much less adopted and adhered to than they should be. Adequate reports of research should describe clearly which questions were addressed and why, what was done, what was shown, and what the findings mean. Our findings will contribute to a set of recommendations to optimize the conduct and reporting of randomized controlled trials and SRs with meta-analysis. Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Background: Reporting guidelines listed at the website of the EQUATOR network, such as CONSORT and PRISMA, are known to improve reliability of medical research literature by promoting transparent and accurate reporting of research studies. Similarly, trial registration improved research transparency and will ultimately strengthen the validity and value of the scientific evidence base. Most biomedical journals require authors to ensure that the manuscript includes all the information recommended in the relevant reporting guidelines, and support clinical trial registration policies by the World Health Organization. However, the attention paid to reporting guidelines and trial registration in nursing journals is unclear. Objectives: The study is aimed to evaluate the editorial policy of nursing journals in the basic requirements for reporting guidelines and trial registration in the last three years. Methods: The ‘Instructions for Authors’ of nursing journals included in ISI Web of Science was reviewed for evidence of an editorial policy on the reporting guidelines and trial registration in March 2012 and March 2015. Results: Of 89 nursing journals examined in 2012, 25 (28.1%) required authors to comply with recommendations in the relevant reporting guidelines, and only seven (7.7%) required clinical trial registration. In 2015, 107 nursing journals were evaluated, 43 (40.2%) required authors to follow the reporting guidelines, and 13 (12.1%) do not consider the trials for publication unless they have been registered prospectively before recruitment of any participants. Conclusions: Three years have passed, most nursing journals still do not require the authors to comply with the recommendation of relevant reporting guidelines, nor claim the need of trial registration prior to the start of patient enrollment in a clinical study. For the purpose of improved reliability in nursing research literature, there is a need for nursing journals to post a requirement on the reporting guidelines and trial registration. P1.057 The information need in systematic review and meta-analysis about acupuncture Wang X1 , Wei D1 , Liu Y1 , Wang Q2 , Qu Q2 , Yang K1 1 Evidence-Based Medicine Center, School of Basic Medical Sciences, Lanzhou University, Key Laboratory of Evidence-Based Medicine and Knowledge Translation of Gansu Province; Chinese GRADE Center, China; 2 School of Basic Medical Sciences, Lanzhou University, China Cochrane Database Syst Rev Suppl 1–327 (2015) 187 Background: As acupuncture becomes more and more popular around the world, the amount of research on it, both primary studies and systematic reviews/metaanalysis(SR/MAs), is increasing rapidly. However, the reporting quality of SR/MAs is poor. The Standards for Reporting Interventions in Clinical Trials of Acupuncture (STRICTA) have shown positive work for reporting quality of clinical trials of acupuncture. Thus, it might be useful to develop an extension of PRISMA for acupuncture to improve the reporting quality of SR/MAs of acupuncture. Objectives: Our first step was to investigate the need for information in SR/MAs of acupuncture from the perspectives of clinicians, researchers, master students and doctors. Methods: Based on a literature review and the PRISMA statement, we designed a questionnaire, containing three parts: the demographic information of respondents; the experience of respondents; the importance of the proposed items. We selected five cities from south, north, north-west and south-west China, to conduct our survey in person, and interviewed those who major in Chinese acupuncture with more than one year’s experience. Results: A total of 269 questionnaires were collected in 18 hospitals, medical universities and research agencies, and 251 (93%) with complete data were used for analysis. The average age of respondents was 33 years. The length of most respondents’ service was less than five years and they read between one and five relevant articles per month. Various approaches for obtaining important information were used by interviewees, and electronic databases, search engines and academic conference were the most used. The most popular types of literature were RCTs (67%), observational studies (51%), basic research (46%), and SR/MAs (37%). Fifty-six per cent of the respondents indicated poor satisfaction with the completeness of information in the literature. Conclusions: Clinicians, researchers and students have a pressing need for complete and critical information from SR/MAs in acupuncture. P1.058 How many journals adopt reporting guidance for RCTs, systematic reviews and Clinical Practice Guidelines in ‘Instructions for authors’? A cross-section survey from 150 medical journals Yao L1 , Chen Y1 , Fu S1 , Wang X1 , Wang Q1 , Wei D1 1 Evidence-Based Medicine Center, School of Basic Medical Sciences, Lanzhou University, China Background: Empirical evidence suggests that active implementation of the CONSORT and other reporting guidance by journals can lead to improvements in the reporting of trials and other studies. Objectives: To investigate how many medical journals adopt reporting standards for randomised controlled trials (RCTs), systematic reviews (SRs) and Clinical Practice Guidelines (CPGs) in their ‘Instructions for authors’. Methods: We used highly sensitive Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. search filters were used to identify RCTs, SRs and CPGs in PubMed. Then we selected the top 50 journals that published each type of study and checked whether CONSORT, PRISMA and reporting criteria for CPGs were adopted in their instructions for authors. Results: The results showed that 28 (56%) journals adopted CONSORT, 20 (40%) journals adopted PRISMA. For CPGs, only one (2%) journal had adopted Conference on Guideline Standardization (COGS) as a reporting standard. Apart from that, two journals mentioned in their instructions that when authors develop guidelines, they should follow Appraisal of Guidelines for Research and Evaluation (AGREE) and the Institute of Medicine (IOM) standard. Conclusions: We suggest that medical journals that publish CPGs should introduce relevant reporting standards in their ‘Instructions for authors’. Guideline authors also need to report their guidelines clearly, transparently and standardly. P1.060 Reports on animal experiments involving neoplasms in Chinese Journals: adherence to ARRIVE guidelines Liu Y1 , Zhao X2 , Mai Y2 , Li X2 , Wang J2 , Chen L2 , Mu J2 , Jin G2 , Gou H2 , Sun W2 , Feng Y2 1 Evidence-Based Medicine Center, School of Basic Medical Sciences, Lanzhou University, Key Laboratory of Evidence-Based Medicine and Knowledge Translation of Gansu Province, China; 2 Evidence-Based Medicine Center, Lanzhou University, China Background: The Animals in Research: Reporting In Vivo Experiments (ARRIVE) guidelines were published in 2010 with the aim of improving the quality of animal study reports. It is not clear how well animal experiments on neoplasms in China adhere to these reporting standards. Objectives: To evaluate the reporting quality of animal experiments involving neoplasms that were published between 2010 and 2012 in Chinese journals and were supported by the National Natural Science Foundation of China (NSFC). Methods: A literature search of studies published between 2010 and 2012 was performed using the Chinese Science Citation Database and the Chinese Journal Full-Text Database. Data were extracted into pre-prepared forms. Reporting quality was assessed using the ARRIVE checklist (40 (39+1) items). Results: A total of 396 animal studies were included in the analysis: 127 published in 2010, 140 published in 2011, and 129 published in 2012. The range of ARRIVE score was 11 to 32. The studies published in 2012 have a greater average ARRIVE score than those published in 2010 (P value 0.034). There was no significant difference between the 2010 and 2011 ARRIVE scores (P value 0.452) or the 2011 and 2012 scores (P value 0.154). Conclusions: Animal experiments of neoplasms in published in Chinese journals did not comprehensively report the information recommended by the ARRIVE guidelines. We strongly recommend researchers report this information when reporting animal studies. Cochrane Database Syst Rev Suppl 1–327 (2015) 188 P1.061 The reporting of items generation and consensus in reporting guidelines: a cross sectional study Chen Y1 , Wang X1 , Wang Q1 , Yao L1 , Wei D1 , Wang M1 , Yang K1 1 Lanzhou University/Chinese GRADE Center, China Background: Empirical evidence suggests that active implementation of the CONSORT and other reporting guidelines by journals can promote improvements in the reporting of trials and other studies. Objectives: To investigate how experts generate items and reach consensus when they develop reporting guidelines. Methods: We searched the EQUATOR library (equator-network.org) on 20 March 2014. Each guideline was reviewed by two independent authors. We only included reporting guidelines on clinical research and systematic reviews. We excluded extension versions and older versions. Results: We identified more than 200 reporting guidelines, but finally 18 met our criteria. The results showed that 14 (78%) described methods of reporting items generation, and 10 (56%) of them adopted more than one approach. The top three methods were literature reviews (13; 72%), adapted existing reporting guidelines (9; 50%), and surveys and interviews (3; 17%). For items consensus, 13 (72%) described methods, 12 (67%) of them mainly used consensus meeting, and four (22%) of them used modified Delphi. Conclusions: Most reporting guidelines described the methods about items generation and consensus in their reporting guidelines, but few of them detailed the processes and implementation. Meta-analysis methods P1.062 The Peto odds ratio viewed as a new effect measure: derivation and application Brockhaus AC1 , Bender R1 , Skipka G1 1 Institute for Quality and Efficiency in Health Care (IQWiG), Germany Background: Meta-analysis has generally been accepted as a fundamental tool for combining effect estimates from several studies within the framework of systematic reviews. In the case of binary data and rare events, the Peto odds ratio (POR) method has become the relative effect estimator of choice. However, as shown in several simulation studies, the POR leads to biased estimates for the odds ratio (OR) when treatment effects are large or the group size ratio is not balanced. This leads to the hypothesis that the POR estimate does not converge towards the true OR even for rare events. Objectives: To investigate the limit of the POR estimator for increasing sample size and its deviation from the OR. Methods: We derived the limit of the POR estimator for increasing sample size by application of the delta method. We investigated in which data situations the POR limit is Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. sufficiently close to the true OR. Results: It was found that the derived limit of the expected POR is not equivalent to the OR, because it depends on the group size ratio. Thus, the POR represents a different effect measure. We investigated in which situations the POR is reasonably close to the OR and derived the maximum effect size of the POR for different group size ratios and tolerated amounts of bias, for which the POR method results in an acceptable estimator of the OR. Conclusions: The POR can be considered as a new effect measure. The POR method can be used as a valid estimate of the OR only in in the presented situations. P1.063 Who should benefit from screening? Are we using the right denominator? Scheibler F1 , Rummer A1 , Sauerland S1 , Grosselfinger R1 1 Institut für Qualität und Wirtschaftlichkeit im Gesundheitswesen (IQWiG), Germany Background: Many screening questions face the problem that, due to low disease prevalence, studies need to have enormous sample sizes in order to detect statistically significant effects on patient-relevant outcomes. Methods: We report on the results of an IQWiG benefit assessment of neonatal pulse oximetry screening (POS) for detection of critical congenital heart disease (cCHD). We compared two different analytical strategies to assess the benefits of POS: the ‘classical’ intention-to-screen (ITS) analysis using all participants as denominator and the ‘alternative’ analysis using only those participants who were affected by the disease. Results: Only one concurrent controlled study (de Wahl-Granelli 2009) could be included in the systematic review. Based on all 155,567 newborns (ITS analysis), the study failed to show a statistically significant effect of screening on severe acidosis (OR 0.490 (0.217 to 1.109) P value 0.086). Using only babies with cCHD as denominator (n = 160), the study reported a significant effect on severe acidosis (OR: 0.268 (0.110 to 0.654) P value 0.003). Discussion: In most screening trials, it is useful to analyze results based on the ITS approach, because the screening interventions themselves affect the prevalence of disease (e.g. by detecting clinically insignificant cases). Thus, using only participants with the disease as denominator might introduce bias into the analysis. In the context of cCHD, however, all affected newborns are destined to die when left untreated, so bias is unlikely. Nevertheless, for comprehensive assessment of a screening intervention, harms have to be examined within the total target population. Conclusion: It is obvious that the benefit of screening primarily evolves from the treatment of affected people. Therefore, if bias is unlikely, the demonstrated benefit in diseased people may justify the implementation of the program, even if the effect on the screened population fails to demonstrate significance. Therefore, if possible and reasonable, when evaluating screening interventions an analysis of results using the affected population as denominator seems to answer additional relevant questions. Cochrane Database Syst Rev Suppl 1–327 (2015) 189 Reference P1.065 de Wahl-Granelli. Impact of pulse oximetry screening on the detection of duct dependent congenital heart disease: a Swedish prospective screening study in 39 821 newborns. BMJ 2009; 338:a3037. doi: http://dx.doi.org/ 10.1136/bmj.a3037 (Published 09 January 2009). Treatment benefit cannot be excluded on the basis of mean difference alone P1.064 Assessment of use and scope for use of IPD in a random sample of 100 reviews from the Cochrane Pregnancy and Childbirth group Sudell M1 , Neilson J2 , Tudur-Smith C1 1 University of Liverpool, United Kingdom; 2 Cochrane Pregnancy & Childbirth Group, United Kingdom Background: Individual Patient Data (IPD) meta-analyses have been described as the ‘gold standard’. However, they are still relatively uncommon within the Cochrane Library and there may be increased potential to improve the findings of traditional meta-analyses by including IPD. Furthermore, traditional reviews may be a useful vehicle for highlighting the added value of IPD for future research. Objectives: To assess the current use, and potential for use of IPD in current Cochrane Reviews produced by the Cochrane Pregnancy and Childbirth Group. Methods: A random sample of 100 of the reviews available on the Cochrane Pregnancy and Childbirth website were downloaded. Various aspects of the reviews were recorded including number of included studies, number of comparisons, risks of bias in included studies and number of heterogeneous comparisons. Any reference to current or future use of IPD was recorded, as well as planned and executed subgroup analyses. A recommendation for whether IPD would be valuable was given based on characteristics of the reviews. Results: Ninety of the 100 reviews included at least one study and 10 were empty. IPD was likely or possibly in use in 11 of the 90 non-empty reviews, however the methods were not well discussed. Seeking IPD was recommended for 64 of the non-empty reviews, and recommended subject to advice from a clinician in a further seven reviews. Common reasons for recommending that a review should seek IPD included presence of heterogeneity that could be explored with IPD to overcome potential bias due to selective reporting or incomplete outcome data, and to overcome insufficient data to complete all planned analyses. Conclusions: Our investigation indicates that few Cochrane Pregnancy and Childbirth reviews currently include IPD, but there is potential to overcome, or at least minimise important issues, by including IPD. However, from a clinical perspective it may not always be justified to seek IPD, as the traditional approach using aggregate data may be sufficient to answer the research questions, and more resources, time and statistical expertise are required for IPD meta-analyses. Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Cates C1 1 Airways Group, United Kingdom Background: The GRADE (Grading of Recommendations Assessment, Development and Evaluation) recommendations (for downgrading due to imprecision) encourage Cochrane authors to be ‘confident’ that a treatment is clinically unimportant if the mean difference and 95% confidence interval (95% CI) fall short of the clinical minimally important difference (CMID). Objectives: To demonstrate that a mean difference alone cannot rule out important clinical benefit in a population. Methods: A Cochrane Airways review will be used to present data to show that a mean difference that is clearly less than the CMID can lead to an important shift in the population of those who respond to treatment (also known as ‘responders’ who individually exceed the CMID). A responder analysis will be compared to the results of modelling the shift in the population mean. Results: A statistically significant increase in the proportion of responders cannot be ruled out by a mean difference and 95% CI that are entirely below the CMID. Conclusions: It is wrong to rule out a treatment as clinically unimportant on the basis of the mean difference alone. If the mean difference shows a statistically significant benefit (however small that might be), then a responder analyses is required to assess the size of the benefit in a population. P1.066 Different mortality time-points in critical care trials: current practice and influence on effect estimates Roth D1 , Heidinger B1 , Havel C1 , Arrich J1 , Gamper G2 , Herkner H1 1 Medical University of Vienna, Department of Emergency Medicine, Austria; 2 Universitätsklinikum St. Pölten, Austria Background: Mortality is frequently used as a primary outcome in critical care trials. It is a patient-orientated variable and robust against information bias. Mortality incidence however needs to be measured at a defined time-point. Practice of meta-analysis in critical care shows that follow-up times of trials in critical care medicine may differ substantially. This has substantial implications on the potential pooling of such mortality estimates. Objectives: Describe the current practice of mortality follow-up time definitions in a representative sample of published critical care randomised controlled trials (RCTs) and analyze the influence of handling of different follow-up times on pooled effect estimates. Methods: We searched CENTRAL, EMBASE, MEDLINE, PASCAL Biomed, and PsycINFO for studies published after 2000 using the Cochrane RCT-filter and a critical care-filter. A random sample of 50% was drawn for further title and abstract review. Study characteristics such as sample size, type of intervention, disease spectrum, intensive Cochrane Database Syst Rev Suppl 1–327 (2015) 190 care unit setting, hospital setting, funding and patient characteristics including age, gender, severity of illness scores were extracted, as well as number and time-points of mortality ascertainment within individual studies. Meta-regression and multilevel mixed-effects linear regression was used to analyze the influence of follow-up time on deviation of pooled risk ratios from study-baseline, using the aforementioned study-characteristics as co-variables, allowing for clustering on level of study and time-point. Results: Search resulted in 9246 studies, 4573 of which were chosen as a random sample. After title-, abstract- and full text-review, 106 studies representing 63,713 patients were included; 60 (57%) studies reported only one time-point, 26 (25%) reported two, 14 (13%) reported three, 4 (4%) reported four, and 2 (2%) reported five. No influence of time-point on effect estimates was found using the aforementioned methods. Conclusions: In a large sample of critical care RCTs, almost half the studies reported more than one mortality time-point. We found no influence of different time-points on pooled effect estimates. P1.067 Results in systematic reviews of anaesthesia interventions that are reported as statistically significant often contain risk of type I error greater than 5% Imberger G1 , Gluud C1 , Boylan J2 , Wetterslev J1 1 Copenhagen Trial Unit, Centre for Clinical Intervention Research, Rigshospitalet, Copenhagen University Hospital, Denmark; 2 Department of Anaesthesia, St. Vincent’s University Hospital, Dublin, Ireland Background: TheGRADEWorkingGroupprovidesanapproach toassessingthequalityofabodyofevidencethatisincreasingly being used, involving explicit consideration of all sources of uncertainty. One of these sources is imprecision, or random error. Many currently published meta-analyses are underpowered and likely to be updated in the future. When data are sparse and there are repeated updates, the risk of random error is increased. Trial sequential analysis (TSA) is one of several methodologies that estimates this increased risk in meta-analyses. With nominally statistically significant meta-analyses of anaesthesia interventions, we used TSA to estimate power and imprecision in the context of sparse data and repeated updates. Methods: We searched for systematic reviews with meta-analyses that investigated anaesthesia interventions.We randomly selected 50 meta-analyses that reported a statistically significant dichotomous outcome in their abstract. We applied TSA to these meta-analyses, using two main TSA approaches: relative risk reduction (RRR) 20% and RRR corresponding to the border of the conventional 95% confidence interval (CI) closest to the null. We calculated the power achieved by each included meta-analysis, using each TSA approach, and we calculated the proportion that maintained statistical significance when allowing for sparse data and repeated updates. Results: In our sample, the median number of trials included was eight (interquartile range (IQR) 5 to 14), the median number of participants was Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. 964 (IQR 523 to 1736) and the median number of events was 202 (IQR 96 to 443). Using both of our main TSA approaches, only 12% (95% CI 5% to 25%) of the meta-analyses had power ≥ 80%, and only 32% (95% CI 20% to 47%) of the meta-analyses preserved the risk of type I below 5%. Conclusion: The majority of nominally statistically significant meta-analyses of anesthesia interventions are underpowered, and a large proportion do not maintain their risk of type 1 error below 5% if TSA monitoring boundaries are applied. A consideration of the effect of sparse data and repeated updates is needed when assessing imprecision in anaesthesia meta-analyses. P1.068 False positive findings in cumulative meta-analysis with and without application of trial sequential analysis Imberger G1 , Thorlund K1 , Gluud C1 , Wetterslev J1 1 Copenhagen Trial Unit, Centre for Clinical Intervention Research, Copenhagen University Hospital, Denmark Background: The majority of meta-analyses are under-powered. The risk of type 1 error in early meta-analyses is higher than the conventional 95% confidence interval (CI) and the associated probability of 5% suggests. Several techniques have been used with the goal of accounting for increased random type I error in the context of sparse data and repeated updates in meta-analysis. This study aimed to explore the role of one of these techniques –trial sequential analysis (TSA) –in assessing the reliability of conclusions in early meta-analyses. Methods: We screened the Cochrane Database of Systematic Reviews and selected 100 meta-analyses that were large enough to have demonstrated, to a reasonable level, that the given intervention does not cause a clinically relevant effect on the outcome in question. We conducted retrospective cumulative meta-analysis using conventional techniques and measured the proportion of false positives that would have occurred. For these false positives, we performed TSA, mimicking how a prospective analysis could have been performed had it been done at the time of publication of each new trial. We used three different TSA approaches to mimic this prospective analysis. As a post-hoc analysis, we surveyed three years of Cochrane Systematic Reviews and calculated the relative risk of a meta-analysis being updated if it was not significant relative to if it were significant. Results: Using conventional retrospective cumulative meta-analysis, one or more false positives were present in seven of the meta-analyses (7%; 95% CI 3% to 14%). Using the three TSA approaches, TSA prevented the false positive type 1 error in 13 of the 14 times the conventional threshold was crossed (93%, 95% CI 64% to 100%). Having a non-significant result made a meta-analysis from a Cochrane Systematic Review in the years 2005 to 2007 1.57 times more likely to be updated (95% CI 0.92 to 2.68). Conclusion: TSA is a helpful statistical methodology when assessing the reliability of early nominally statistically significant findings in cumulative meta-analyses. Cochrane Database Syst Rev Suppl 1–327 (2015) 191 P1.069 P1.070 Proving clinical relevance of effects in meta-analyses of binary data From meta-analysis to practical decision using virtual realistic population simulation: an illustration on sudden death prevention in type 2 diabetes Skipka G1 , Bender R1 , Lange S1 1 Institute for Quality and Efficiency in Health Care (IQWiG), Germany Background: Often, it is not sufficient to prove statistical significance of effects. Rather, the clinical relevance of effects has to be shown. Usually, clinical trials are powered for statistical significance only. Within the framework of systematic reviews the power (precision) is increased by pooling results of multiple trials. Therefore, more stringent conclusions towards clinical relevance can be drawn without loss of power. Objectives: To develop an approach to prove clinical relevance in meta-analyses of binary data. Methods: The starting point is a specified true effect that is considered clinically relevant. Also in order to prove clinical relevance, the probability of statistical errors should be controlled. Therefore, the precision of the estimates has to be considered. This is fulfilled by determining a threshold for the confidence interval (CI) obtained by meta-analysis which corresponds to the testing of shifted hypotheses. This threshold has to be precisely selected so that a specified power is preserved. The main idea is that the power of the meta-analysis should be identical to the power for proving statistical significance in the individual studies included in the meta-analysis. Balanced sample sizes are assumed. This approach will be investigated for the effect measures relative risk (RR) and odds ratio (OR). Results: Simulations show that the resulting thresholds depend on the baseline risk to some extent. The OR is substantially more severely affected than the RR. Therefore, the RR is favoured. In addition, the baseline risk dependency nearly vanishes if using the restricted maximum likelihood approach (REML) instead of the Wald approach for the calculation of the CIs. Conclusions: Clinical relevance can be shown without loss of power by taking advantage of the higher precision of meta-analyses. The thresholds for clinical relevance are nearly independent of the baseline risk if the REML-approach is applied. Thus, the determination of these thresholds only depends on the higher precision gained by meta-analyses. Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Le H1 , Marchant I2 , Boissel J1 , Nony P1 , Kassaı̈ B1 , Cornu C1 , Gueyffier F1 1 UMR 5558, Université de Claude Bernard Lyon 1, France; 2 Universidad de Valparaı́ so, Chile Background: Meta-analysis in reviews systematically results from numerous randomized controlled trials and supplies synthesis information for clinical practice. However, there is no standard procedure to predict the impact of translating these results at a population level. Objectives: We propose a strategy to simulate public health impact from the results of meta-analyses combined with a risk score applied on a virtual realistic population (VRP) platform. We simulated a preventive strategy for sudden death (SD), a major cardiovascular event, in a French population with type 2 diabetes (T2D). We illustrated how this approach could help to rationalize health public decisions. Methods: - Generate a French diabetic VRP. - Establish a SD risk score. - Review existing meta-analyses on therapeutic efficacy in preventing SD in T2D. - Apply these results on the generated VRP with drug combination and risk-based patient selection. Results: A French diabetic VRP of 176,187 was generated from a 8995-patient sample, giving an median of SD risk of 1.7% at a 5-year time horizon. We estimated the numbers needed to treat (NNTs) by a simultaneous treatment by statin and ACE inhibitor at 221 people for the whole population, and at 108/104 women/men among individuals of the highest 10% predicted SD risk. The corresponding untreated risks of SD were 1.9% on the whole population, and 3.9% and 4.1% in top risk deciles for women and men respectively (Table 1). Conclusions: Our approach, which gathers effect models (via meta-analyses and risk scores) and VRP simulation, provides a powerful multi-component tool for valuing each evidence-based component, better transposing clinical trial results to practice, and facilitating clinical decision in both public health and individual levels, on both medico-economic aspects. Key words: health public simulation; meta-analysis; practical decision making; sudden death; virtual realistic population. Cochrane Database Syst Rev Suppl 1–327 (2015) 192 P1.072 A network meta-analysis of randomized controlled trials of interventions for the treatment of rotator cuff tears Gagnier J1 , Beyene J2 1 University of Michigan, USA; 2 McMaster University, Canada Background: Rotator cuff tears are common and increasing in prevalence in our aging population. Rotator cuff tears can be treated through various surgical procedures and rehabilitative programs, many of which have not been compared directly to each other. Using a network meta-analysis allows comparison and facilitates pooling of a diverse population of randomized trials across these approaches in ways that a traditional meta-analysis does not. Objectives: Our aim is to perform a network meta-analysis using evidence from randomized trials on the relative effect of alternative approaches on recovery in patients who have suffered from a rotator cuff tear. Our secondary study endpoints included a repeat tear. Methods: A network meta-analysis allows for simultaneous consideration of the relative effectiveness of multiple treatment alternatives. We began with systematic searches of databases (including EMBASE and MEDLINE) and performed hand searches of orthopaedic journals, bibliographies, Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. abstracts from orthopaedic conferences, and orthopaedic textbooks, for all relevant material published from each databases’ inception to 2015. Two authors independently screened abstracts and manuscripts and extracted the data, three evaluated the risk of bias in individual studies, and two applied Grading of Recommendation Assessment, Development and Evaluation (GRADE) criteria to bodies of evidence. We included all randomized and quasi-randomized trials comparing two (or more) treatment options for rotator cuff tears in predominantly (i.e. > 80%) adult patients. We calculated pooled estimates for all direct comparisons and conducted a network meta-analysis combining direct and indirect evidence for all comparisons. Results: We included 35 papers comparing two or more interventions for rotator cuff tears. A total of four interventions were tested across trials. We are currently performing the statistical analyses for the network comparisons. Conclusions: This will be the first network meta-analysis done for clinical trials of rotator cuff tears. These findings will help guide clinical decision making and the design of future randomized studies. Cochrane Database Syst Rev Suppl 1–327 (2015) 193 P1.073 Exploring the efficacy effectiveness gap Koesters M1 , Holtrup A1 , Fiedler I1 , Becker T1 1 Department of Psychiatry II, Ulm University, Germany Background: There is broad consensus that there is a gap between the ‘efficacy’ of a drug proven in randomized controlled trials (RCTs) and the ‘effectiveness’ of that drug in everyday clinical practice. This consensus is mainly based on anecdotal evidence. Objectives: This study systematically reviewed and compared evidence from RCTs and nonrandomized and uncontrolled trials examining the effectiveness of venlafaxine and duloxetine in the treatment of depression. Methods: The systematic review used pre-post effect sizes to compare treatment effects from RCTs and non-RCTs examining the effects of duloxetine, venlafaxine and placebo in the treatment of depression. Data were aggregated in random-effects models. Meta-regressions were used to explore factors influencing the effect sizes. Results: One-hundred and ninety relevant studies were identified, and effect sizes for 114 acute-phase studies could be calculated. All interventions, including placebo, showed a statistically significant pre-post effect of more than two standard deviations. Between-study heterogeneity was high. However, there were no significant differences of effect sizes from RCTs and non-RCTs, and metaregressions did not show a significant influence of external validity measured by the Downs and Black scale. Conclusions: The analyses could not confirm the presence of an efficacy-effectiveness gap.It is possible that the impression of such a gap existing is a consequence of publication bias rather than being due to RCT-practice differences. P1.074 Application of Bayesian methods in public health systematic reviews Lewis MG1 , Nair S2 1 PhD student, India; 2 Director Public Health Evidence South Asia (PHESA), India Background: Bayesian meta-analysis has gained popularity in the field of evidence synthesis. Public health interventions are targeted to highly heterogeneous populations, multi-component interventions, multiple outcomes, influenced by context and, moreover, the answer to a question comes from a hierarchy of studies, which is in contrast to the clinical interventions. In such situation Bayesian methods have a wide scope of application. Objectives: To develop methods to incorporate Bayesian methods into public health evidence synthesis. Methods: Public health systematic reviews were identified which included studies with different type of study designs for a single question. Data were extracted from those studies for the outcome(s) concerned. Bayesian meta-analysis was performed with likelihood from the stronger studies and the Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. results of the weaker studies as prior. The analysis was performed in WINBUGS software with 5000 iterations for burning the sampler and 10,000 iterations for estimation. Convergence was assessed using autocorrelation and density plot. A simulation study was also performed with different meta-analysis situations with repetitions of 1000 in each case. Results: The concept has been applied in ten public health systematic reviews, which included different study designs to answer the same question. Different combinations of priors and likelihood were drawn and odds ratio, relative risk with 95% CrI, CI were calculated. The results were then compared with the estimates of traditional meta-analysis with the posterior estimates of Bayesian. Conclusions: The present paper describes a mechanism to incorporate all the levels of a body of evidence for a single question using the Bayesian approach. P1.075 An empirical study investigating the extent of heterogeneity, inconsistency and potential limitations in networks of interventions published in 2013 Gianatsi M1 , Nikolakopoulou A1 , Chaimani A1 , Salanti G1 1 Department of Hygiene and Epidemiology, School of Medicine, University of Ioannina, Greece Background: Network meta-analysis (NMA) has become an increasingly popular methodology and several Cochrane Reviews comparing multiple interventions have been published over the last years. Despite the rapid methodological development of NMA, the good practice requirements of this advanced evidence synthesis tool are often ignored or improperly applied. Objectives: To update the current empirical evidence on characteristics of published NMAs, including their methodological rigor and quality of reporting, and to evaluate empirically the extent of heterogeneity and inconsistency in networks of interventions. Methods: We searched MEDLINE, the Cochrane Library and EMBASE to identify NMAs published in 2013. We extracted data on various network characteristics and assessed how often researchers evaluated and reported the assumptions underlying NMA appropriately. We documented the methods employed to estimate and present relative effects and treatment ranking and assessed the extent to which authors discussed the potential for bias in their findings. We estimated three indices that quantify the percentage of variability attributed to heterogeneity, inconsistency, or both, for each primary outcome. Results: Our search identified 72 networks that met all inclusion criteria. So far we have extracted data from 40 networks: only 25% discussed how plausible the assumption of NMA is. Inconsistency was assessed statistically in less than half of the networks (43%), while issues of bias (e.g. publication bias or attrition bias) were discussed in about one in six networks. The I-squared statistic for heterogeneity or inconsistency ranged between 25% and 75% in almost half of the networks. Results from the 72 Cochrane Database Syst Rev Suppl 1–327 (2015) 194 networks will be presented at the Colloquium. Conclusions: Assumptions and potential limitations of NMA are not widely discussed in publications and this casts doubts about the credibility of their conclusions. The majority of networks in our database presented a moderate level of heterogeneity and inconsistency. Siersma V, Als-Nielsen B, Chen W, Hilden J, Gluud LL, Gluud C. Multivariable modelling for meta-epidemiological assessment of the association between trial quality and treatment effects estimated in randomized clinical trials. Stat Med 2007;26:2745–58. P1.077 P1.076 metaepi: an R package for conducting meta-epidemiologic studies Eusebi P1 , Abraha I1 , Cozzolino F1 , Siersma V2 1 Regional Health Authority of Umbria, Italy; Copenhagen, Denmark 2 University of Background: Various factors in the conduct of clinical trials may bias the intervention effect. On the basis of theoretical and empirical considerations, Cochrane identifies several factors to be assessed for risk of bias in randomized trials. Empirical evidence comes from meta-epidemiology, where the intervention effects from the studies in a collection of meta-analyses are compared between trials with and without a particular characteristic. Meta-epidemiological studies can be performed with a two-step approach (Sterne 2002) or a one-step approach (Siersma 2007). Investigation of heterogeneity, sensitivity analyses and graphics are key points. Outcomes can be continuous or discrete. Objectives: To develop an R package for standardizing the procedure of a meta-epidemiologic study. Methods: Several functions are developed in the R package allowing for: (1) fitting twoand one-step models; (2) performing meta-regression and subgroup analyses; (3) drawing Baujat plots, forest plots and funnel plots. The package’s functions are illustrated on data from a meta-epidemiologic study investigating the impact of deviations from intention-to-treat on the intervention effect (provisionally accepted for publication). Results: The functions of the package deploy the R utilities in manipulating data, the lme4 package for mixed-effects modelling and the base graphics system. The functions are easy to use and allow for well-structured meta-epidemiologic analysis. Conclusions: The R package metaepi is a convenient tool for facilitating the statistical analysis in a meta-epidemiologic study and can provide guidance for the steps required in performing such challenging task. Reference Sterne JA, Juni P, Schulz KF, Altman DG, Bartlett C, Egger M. Statistical methods for assessing the influence of study characteristics on treatment effects in ‘meta-epidemiological’ research. Stat Med 2002;21: 1513–24. Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Similarity of the treatment benefit between children and adults: results of a meta-epidemiological study Janiaud P1 , Lajoinie A1 , Cour-Andlaueur F2 , Cornu C1 , Cochat P3 , Cucherat M1 , Gueyffier F1 , Kassai B2 1 University of Lyon, France; 2 EPICIME-Clinical Investigation Center, INSERM CIC201/UMR5558, France; 3 CHU Lyon, Service de Pharmacologie Clinique, France Background: Therapeutic benefit is commonly extrapolated from adult clinical trials to children without conducting trials in children. It has been widely recognized that children cannot be provided with safe and efficacious drugs compared to those available for adults without involving them in clinical trials. To our knowledge, extrapolation of data from adults to children based on the results of randomized controlled trials (RCTs) has not consistently been explored. Objectives: Our main objective was to see whether the therapeutic benefit observed in placebo controlled RCTs is different between adults and children. Methods: We searched three electronic databases for meta-analyses that included double-blind, placebo–controlled RCTs with separate results for adults and children. The selected reviews were classified according to disease and drug used. The heterogeneity of treatment response between adults and children was measured using ratio of odds ratios (RORs). Results: We selected 89 meta-analyses and calculated RORs for 124 drugs. Heterogeneity in the direction of the treatment effect was observed in one drug and heterogeneity in the quantity of the treatment effect in 13 drugs (Figure 1), indicating significantly different treatment effect in adults when compared with children. RORs were not significantly different from 1 for 110 drugs. For 36 of these drugs, the treatment effect was confirmed in both populations. Conclusions: We found different treatment benefits estimated by clinical trials performed in adults compared with those performed in children for 14 out of 124 drugs. Our results also show the potential influence of pharmacokinetics and pharmacodynamics and the need to better report on the rationale of dose adjustments in RCTs. A better reporting of age range and dose adjustment would allow full exploration of possible sources of clinical and statistical heterogeneity, enabling the prediction of situations where there may be differences between adults and children. Cochrane Database Syst Rev Suppl 1–327 (2015) 195 Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Cochrane Database Syst Rev Suppl 1–327 (2015) 196 P1.078 Similarity of the benefit risk ratio between children and adults: results of a meta-epidemiological study Janiaud P1 , Chiganne M1 , Hergibo F1 , Araujo I1 , Kassai B2 1 University of Lyon, France; 2 EPICIME-Clinical Investigation Center, INSERM 7 CIC201/UMR5558, France Background: Adverse events (AEs) are a major cause of morbidity and pose a substantial burden on limited healthcare resources. Drug efficacy and safety data are often extrapolated from adults to children. Children can be more resilient to AEs or more vulnerable. Objectives: Our objective was to assess whether the benefit-risk ratio is different between adults and children using evidence from placebo-controlled randomized clinical trials (RCTs). Methods: We searched three electronic databases for meta-analyses that included double-blind, placebo-controlled RCTs with separate results for adults and children. For the quantitative synthesis only meta-analyses that had an efficacy primary outcome and at least one safety outcome (adverse events (AEs) and/or withdrawals due to AEs) were retained. The net efficacy adjusted for risk odds ratio (NEAR-OR) was calculated for Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. adults and children using the expected frequency of patients responding without harms. The best benefit risk profile corresponds to the greatest number of patients to respond favorably without suffering adverse events. The heterogeneity of the benefit risk ratio between adults and children was measured using the ratio of NEAR-ORs. Results: Only 20 out of 89 included meta-analyses, evaluating 25 drugs, had data allowing the quantitative synthesis. For two drugs the NEAR-OR changed direction compared to the efficacy OR in children (Figure 1). For example for lamotrigine for drug resistant epilepsy, the OR showed a significant treatment benefit in children but the NEAR-OR showed significantly more harms. For AEs outcomes, two NEAR-RORs were significantly different from 1 (Figure 2). One drug showed a higher treatment benefit without AEs for children and the other showed a higher benefit risk ratio in adults. Conclusions: We found a difference in the benefit risk ratio between adults and children for two drugs. The adjustment of the treatment benefit on risks led to a loss of benefit for children for two drugs. The NEAR-OR is a new method which seems useful when assessing the benefit risk ratio of drugs. It needs to be more widely used in order to be validated and to provide robust conclusions. Cochrane Database Syst Rev Suppl 1–327 (2015) 197 Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Cochrane Database Syst Rev Suppl 1–327 (2015) 198 P1.080 The waste of research in the addiction field: an example from a review on psychological intervention for psychostimulant-use disorders Minozzi S1 , Mitrova Z1 , Saulle R1 , De Crescenzo F2 , Amato L1 1 Cochrane Drugs and Alcohol Group, Italy; 2 Institute of Psychiatry and Psychology Catholic University of the Sacred Heart, Rome, Italy Background: There are a fair number of reasons that have been explored for waste in research, such as when the needs of stakeholders are ignored, or when what is already known is Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. overlooked. We would highlight waste of research, due to the great variety of interventions compared without consideration of the existing evidence. Objectives: To assess the number of comparisons for which meta-analyses can be done, and to quantify the waste of studies that have to be excluded based on the variety of the comparisons in the field of psychostimulant addiction. Methods: Data will be extracted from an ongoing Cochrane Systematic Review assessing the effectiveness of psychological intervention for psychostimulant-use disorders. Experimental and control interventions will be listed as well as different types of comparisons performed. The number of studies with comparisons similar enough to be pooled in meta-analysis will be ascertained, as well as the types of comparisons that will be reported finally in the ‘Summary Cochrane Database Syst Rev Suppl 1–327 (2015) 199 of findings’ table. The number of studies excluded will be quantified and the types of comparisons wasted will be described. Results: Preliminary results based on 93 studies included showed that 15 comparisons can be done, with a median number of three studies (range 1 to 38). Three comparisons included only one study and another three only two studies. High variability in outcomes considered represents an additional limit to synthesizing results. Conclusions: Our findings could help to highlight a major issue of relevance of primary research in the field of addiction, giving a hint in the debate on reasons for waste of research and the potential role of Cochrane Reviews in informing the research agenda. P1.081 Methods for individual participant data meta-analysis of relative treatment effects: a methodologic overview Debray T1 , Moons K1 , van Valkenhoef G2 , Efthimiou O3 , Hummel N4 , Groenwold R1 , Reitsma H1 1 Julius Center for Health Sciences and Primary Care, The Netherlands; 2 University Medical Center Groningen, The Netherlands; 3 University of Ioannina, Greece; 4 Universität Bern, Switzerland Background: Individual participant data (IPD) meta-analysis is an increasingly used approach for synthesizing and investigating treatment effect estimates. Over the past few years, numerous methods for conducting an IPD meta-analysis (IPD-MA) have been proposed, often making different assumptions and modeling choices while addressing a similar research question. As a consequence, meta-analytic models sometimes lead to conflicting results, which casts doubt on their utility and validity. Objectives: To assist researchers to choose appropriate methods and provide recommendations on their implementation when planning and conducting an IPD-MA. Methods: We conducted a literature review to provide an overview of methods for performing an IPD-MA using evidence from clinical trials or non-randomized studies when investigating treatment efficacy. Articles were eligible if they presented statistical methods, results from empirical case studies or provided methodological guidance. Results: A total of 153 records (out of 3360 unique records identified) were included. Conclusions: IPD-MA offer numerous advantages over meta-analyses that are solely based on published AD. There is substantial, although disperse, information on how IPD-MA should be conducted. This review may help systematic reviewers with a limited background in medical statistics to identify relevant methods for their IPD-MA. Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. P1.082 Acupuncture therapy for diabetic peripheral neuropathy: a network meta-analysis and methodology research Xiong W1 , Chen W1 1 Evidence-Based Chinese Medicine, Beijing University of Chinese Medicine, China Background: Well-designed multicenter randomized controlled clinical trials (RCTs), systematic reviews and meta-analysis could just assess two interventions treating the same disease. However, when there are more than two treatments for the same disease, direct comparisons between each treatment should not be forgotten. Network meta-analysis (NMA) aims to compare and estimate the pair wise effect sizes of variety of treatments for the same disease simultaneously when there is a lack of sufficient evidence from direct comparisons. Objectives: To evaluate the application of NMA in the therapeutic evaluation of acupuncture therapy for the treatment of diabetic peripheral neuropathy (DPN), and to apply NMA to acupuncture therapy and traditional Chinese medicine. Methods: We searched the following databases to September 2014: the Cochrane Central Register of Controlled Trials (CENTRAL), PubMed, CNKI, VIP, Wanfang, and SinoMed. After including RCTs of acupuncture therapy for DPN, we performed NMA and drew network plots through GeMTC R package and WinBUGS software based on the Bayesian statistical model and Markov Chain Monte Carlo method (MCMC). Results: Our research involved seven treatments (manual acupuncture, electro-acupuncture, tapping, needle warming moxibustion, mecobalamin, no treatment and vitamin B1 and B12) in 40 trials, and a total of 2602 participants. The network plot drawn by R software through calling the Cochrane Database Syst Rev Suppl 1–327 (2015) 200 GeMTC R package has been appended to the attachments. The Rank probability showed that four types of included acupuncture therapy all had beneficial effects on global improvement, especially tapping, which displayed the best efficacy, and needle warming moxibustion which followed it. However, the poor quality of the methodology may lead to a need to be more cautious when interpreting these positive findings. Conclusions: The results may provide valid certified clinical evidences of acupuncture therapy for the treatment of diabetic peripheral neuropathy as well as the feasible solution for the methodological issues in the NMA, which can help to establish a reference model therapeutic evaluation of traditional Chinese medicine. P1.083 Network meta-analyses in clinical practice guidelines: a cross-sectional survey Yao L1 , Fu S1 , Cui R1 , Deng W1 , Zhang P1 1 Evidence-Based Medicine Center, School of Basic Medical Sciences, Lanzhou University, China Background: The development of clinical practice guidelines (CPGs) should be based on systematic reviews. Network meta-analyses (NMAs) are useful in CPGs when there is no evidence that compares interventions of interest directly (indirect treatment comparison; ITC), or when combining direct and indirect evidence (multiple treatment comparison; MTC). This study is to analyze the proportion of NMAs used in CPGs. Objectives: This study is to analyze the proportion of NMAs used in CPGs. Methods: PubMed database was searched to obtain NMAs; 100 NMAs were randomly selected as the sample by SPSS 19.0, then we collected the articles that cited the NMAs sample in Google and Web of Science. Results: A total of 289 NMAs were retrieved from PubMed, and 100 NMAs were sampled randomly. A total of 2026 articles had cited the NMAs sample, among which 50 (2%) were CPGs, including 12 National Guideline Clearinghouse (NGC) CPGs and two National Institute for Health and Care Excellence (NICE) CPGs. All of the CPGs that cited NMAs were conducted or updated in 2011 to 2013. Conclusions: The value of NMAs could make explicit the decisions made implicitly when multiple pairwise comparisons are presented in CPGs. However a low proportion of NMAs were used in CPGs. Further research on how to improve use of NMAs in CPGs would be valuable. Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. P1.086 Cumulative meta-analysis of the efficacy of acupuncture on conventional therapy-induced side effects in oncology patients Pan Y1 , Shen X2 , Shi X3 , Zhang H3 1 Institute of Medical Psychology, School of Basic Medical Sciences, Lanzhou University, China; 2 Institute of Epidemiology and Health Statistics, School of Public Health, Lanzhou University, China; 3 Department of Internal Medicine, Gansu Rehabilitaition Center Hospital, China Background: Current meta-analyses have investigated the effectiveness of acupuncture for management of conventional therapy-induced side effects of cancer, however, the results were inconsistent. Little is known about the cumulative meta-analysis of acupuncture practice; it is important that the cumulative effect of the various proposals is kept under consideration. Objective: To evaluate empirically the cumulative performance of acupuncture in meta-analysis on therapyinduced side effects of symptoms in cancer patients. Methods: Databases: MEDLINE, EMBASE, PsycLIT, MANTIS, Science Citation Index, Cochrane Controlled Trials Registry, Cochrane Complementary Medicine Field Trials Register, Cochrane Pain, Palliative Care and Supportive Care Specialized Register, Cochrane Cancer Specialized Register and conference abstracts were consulted up to January 2015. We evaluated the quality of the included studies by AMSTAR criteria and analyzed the data using the Stata software, version 10.0. Data were provided by investigators; odds ratios and 95% confidence intervals were calculated on dichotomous data. Results: Five meta-analyses have been published from 2006 to 2013, including a total of 5235 cancer participants and 5685 controls. The effect size and confidence intervals in pooled analysis failed in favor of acupuncture group compared with control therapies (acupuncture versus control groups, odds ratio (OR) = 1.00, 95% confidence interval (CI): 0.87 to 1.33). Furthermore, the confidence intervals of treatment effectiveness of total population can be narrowed by accumulation in chronological order., the impact of sample size on the comprehensive results can be observed by accumulation according the sample size. Conclusions: Cumulative meta-analysis methods can provide more information than the traditional meta-analysis methods. The shortage of meta-analysis may have caused a lack of statistical significance, which could be explained by the small sample size of the current study, acupuncture prescription, primary endpoint, the possibility of heterogeneity still exists, and the findings could not result in a reasonable recommendation. Cochrane Database Syst Rev Suppl 1–327 (2015) 201 Trial registration P1.087 Practical and conceptual issues of trial registration Freitas C1 , Pesavento T1 , Pedrosa M1 , Riera R1 , Torloni MR1 , Nogueira G1 , Porfirio G1 , Silva A1 , Atallah A1 , Logullo P1 1 Brazilian Cochrane Centre, Brazil Background: Trial registration is a prerequisite for publication of studies in respected scientific journals. Recent Brazilian regulations also require the registration of some clinical trials in the national registry (ReBEC), but there is little information available about practical questions involved in the registration process. Objectives: To discuss the importance of trial registration and practical issues involved in this process. Methods: Information was obtained from trial registry platforms, reference lists, and websites (last search: September, 2014) on the following topics: definition of a trial, history, purpose and importance of trial registries, which information should be registered and the process of registration. Results: Clinical trial registration aims to avoid publication bias. In Brazil, it is required by journals indexed in LILACS, SciELO; worldwide it is required by ICMJE-affiliated journals. Recent Brazilian laws require that all trials (phases I to IV) involving new drugs to be marketed in the country must be registered in the Brazilian Clinical Trials Registry (ReBEC). The pros and cons of using different trial registration platforms are discussed in an article accepted by the Sao Paulo Medical Journal. Conclusions: Trial registration is important and currently there are mechanisms to enforce its implementation. Researchers should take into account national laws and publication requirements when choosing the platform on which they will register their trial. P1.088 An audit of the data structures of four international clinical trial registries Ko H1 , Hunter K1 , Vu T1 , Suasa R1 , Smith E1 , Zhang L1 , Askie L1 1 NHMRC Clinical Trials Centre, University of Sydney, Australia Background: Clinical Trial Registries (CTRs) are increasingly used by systematic reviewers to identify ongoing and completed clinical trials. CTRs have trial information on at least 20 unique data items (i.e. WHO Trial Registration Data Set (TRDS)). Whilst being compliant with the TRDS, the content between registries may be inconsistent as CTRs operate independently of each other. This presents difficulties when extracting data from multiple CTRs. Objectives: To characterize the content of the data fields in the 20 TRDS items across different CTRs, and to identify data fields where extracting consistent information was complex. Methods: Data fields from 4 CTRs (ANZCTR, ClinicalTrials.gov, ISRCTN, EuCTR) were audited in Dec 2014 as part of a larger project analysing country-specific clinical trial activity. The 20 TDRS Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. items were assessed for their data quality, and how CTRs varied regarding the number of data items per data field, aggregated data items, and data item formats. Results: All 4 CTRs collected the 20 TRDS items, but the content and number of data fields used to collect these items was not consistent across CTRs. Some TRDS items required single data fields only e.g. registration number. The more information-intensive items often had multiple, associated data items or were aggregated into single data items, e.g. intervention description, sponsor, and funder items differ across CTRs. Data items were presented in different formats e.g. one data item might have a combination of categorical, binary, or free text data. Some non-TDRS data items were in some CTRs but not others. Some CTRs have coding options for certain data items that are not directly translatable in other CTRs, e.g. intervention condition codes. Conclusions: It is a complicated process for users to extract consistent data across multiple CTRs. Due to variations between CTRs (re: content, where information is kept, number of fields, types of data and how information is coded), accurate information extraction requires data field matching, formatting, cleaning, and recoding. Further standardization of CTR data is needed to improve the ease with which CTR data can be more fully utilised. P1.089 Australian New Zealand Clinical Trials Registry (ANZCTR) user survey results from 2009 and 2015 Hunter K1 , Vu T1 , Sausa R1 , Tan-Koay A1 , Ko H1 , Askie L1 1 NHMRC Clinical Trials Centre, University of Sydney, Australia Background: Clinical trial registries are an important resource for researchers, consumers and clinicians. In order to maximise their utility, it is important to obtain user feedback. ANZCTR has conducted two user surveys since its inception in 2005. Objectives: To obtain user feedback on the relevance, functionality and usefulness of the ANZCTR. Methods: There were two survey periods: Sep 2008 to Mar 2009 and Nov 2014 to Mar 2015. Users were prompted to complete a survey upon visiting the website. Both surveys were identical, containing 8 items covering user role, frequency and ease of website use, type of information being sought and relevance of the registry. Users were required to select appropriate option(s) from a list, and a free text item was also included. In the first survey the question about role was mandatory. Results: In the 2008/9 survey 16,356 users responded to the mandatory question about their role, and a smaller proportion (∼2500) completed some/all of the remaining optional questions. In the 2014/15 survey 2662 users responded to at least one question. Both surveys indicated that the most common role was a registrant (28.9%), followed by researcher (26.9%), consumer (19.5%), clinician (10.1%), journal editor/staff (2.6%), funder (1.7%) and regulator (1.1%). Most respondents were first time users of the site (42.4%), sought general information about clinical trials (19.4%), and were able to find what they were looking for most/all of the time (61.2%). Most also agreed/strongly Cochrane Database Syst Rev Suppl 1–327 (2015) 202 agreed that prospective trial registration is a good idea (84.2%) and that ANZCTR is an important national resource (80.2%). A comparison between the two survey periods showed a 5.0% increase in the proportion of researchers using the site and 6.2% increase in the proportion of respondents who strongly agreed that ANZCTR is an important national resource. Conclusions: The most common users of ANZCTR were registrants and researchers. Whilst most respondents were able to find what they were looking for, there is scope for improvement of the search function and navigation interface of the website. The vast majority agreed that ANZCTR is an important national resource. P1.090 Is it necessary to search trials registers? Mann M1 , Weightman A2 , Coles B3 1 Cardiff University, United Kingdom; 2 Cochrane IRMG, United Kingdom; 3 Cancer Research Wales/Cardiff University, United Kingdom Background: In order to comply with the Cochrane Handbook and Methodological Expectations of Cochrane Intervention Reviews (MECIR), Cochrane authors are obliged to search trials registers where relevant to their review topic area. Searching trial registers retrieves ongoing or unpublished studies, which will reduce the risk of publication bias. Objectives: Our study investigated the extent to which Cochrane authors are searching trials registers. Furthermore, to identify whether any of the included unpublished studies were retrieved from a trials register search. Methods: We examined 27 reviews flagged as ‘New’ in The Cochrane Library (Issue 3, 2015) 19 March 2015. We determined reviews where trials registers had been searched. We extracted data to identify whether any included studies in the reviews were found uniquely in a trials register. Results: Twenty reviews (74%) included a search of trials registers. Some reviews searched other registers in addition to the registers suggested in MECIR as mandatory. In our poster presentation, we will fully discuss the findings as to how far registers provide a unique source of unpublished but relevant studies. All review results will be anonymous. Conclusions: Even though endorsement of MECIR recommendations has been in place since 2013, some authors of new reviews were not compliant with these standards. Therefore, Cochrane Review Groups should be more explicit in their expectations of authors about standards for MECIR. P1.091 The World Health Organization’s International Clinical Trials Registry Platform: a ‘one-stop shop’ for identifying clinical trials conducted worldwide? Corp N1 , Jordan J1 1 Research Institute for Primary Care & Health Sciences, Keele University, United Kingdom Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Background: The World Health Organization (WHO) International Clinical Trials Registry Platform (ICTRP) is a central repository for clinical trials regularly updated from 15 primary and three partner registries worldwide, the former meeting requirements of the International Committee of Medical Journal Editors. As such, it has the potential to be a valuable one-stop access point to identify on-going, recently completed and unpublished trials, negating the need for multiple searches across multiple registries. However, whilst the WHO ICTRP includes the Clinicaltrials.gov registry, the current methodological standards for Cochrane Reviews stipulate the mandatory search of both, questioning its functionality and utility. Objectives: This study aims to assess the functionality and utility of the WHO ICTRP as a ‘one-stop shop’ for identifying clinical trials by comparing search facilities and outputs with those from the original primary trials registers. Methods: Using specific musculoskeletal conditions as examples, WHO ICTRP and its 15 primary registries, will be searched individually to identify relevant clinical trials: simple (condition only) and more complex (e.g. stipulating interventions, trial end or registration dates) searches will be undertaken and search outputs compared for agreement. Results and Conclusions: We will present results to clarify whether WHO’s ICTRP offers a valuable ‘one-stop shop’ for retrieving clinical trial information, or whether the additional time implicit in searching the original trials registries is time well spent in terms of acquiring a comprehensive overview of registered trials. We will also present an assessment of the search capabilities, information attainable and handling of search outputs of each of the included registries and discuss the implications for Cochrane Reviews and other systematic reviews. P1.092 The German Clinical Trials Register (DRKS) and its data quality compared to other trial registries Jena S1 , Kunzweiler K1 University Medical Center, Freiburg, Germany 1 Background: Publicly accessible clinical trials registries are crucial for providing information about planned, ongoing, completed or even discontinued clinical trials for patients and their doctors, scientists, health professionals, funding organizations and the general public. National registries such as the German Clinical Trials Register (DRKS) are an indispensable tool to inform the public in the national language. Practitioners, patients and laymen clamour urgently for data available in national languages. These registries help to: - conduct reviews, meta-analyses, and medical guidelines using a basis of comprehensive and complete trial information; - avoid misinformation and to reduce bias in health care and health research; - utilize resources reasonably so that clinical trials don’t have to be reiterated, exposing patients to an unnecessary risk. To meet these needs requires comprehensive, complete, Cochrane Database Syst Rev Suppl 1–327 (2015) 203 and consistent data in clinical trial registries. Publications regarding data quality in clinicaltrials.gov (1) and the World Health Organization (WHO) Meta-Portal (2) showed the data quality is insufficient. Objectives: The trial data in DRKS will be examined analogously for completeness, accurateness and consistency. A lot of automatic and manual queries in DRKS should lead to a high data quality, that stands out from other databases. Methods: Data on the following four topics will be extracted, coded and analyzed: intervention of the trial, endpoints, contact details, study design. For all publically accessible fields the filling degree will be calculated. Results: Preliminary evaluations showed consistent and nearly complete information in all analyzed fields. Further analyses will be presented. Conclusions: High data quality in trial registries is essential for clinical research and health care, DRKS endeavours to contribute to this scientific task. References 1. Zarin DA, Tse T, Williams R et al. The ClinicalTrials.gov results database-update and key issues. N Engl J Med 2011: 364: S852–S860 2. Viergever RF, Ghersi D, Scherer RW. The quality of registration of clinical trials. PLOS ONE 2011: 6: Se14701. P1.093 Current malaria clinical trials activity on the Pan African Clinical Trials Registry and comparison to a historical cohort Lutje V1 , Pienaar E2 , Kredo T2 , Abrams A2 1 Cochrane Infectious Diseases Group, United Kingdom; 2 South African Cochrane Centre, South African Medical Research Council, South Africa Background: Ninety per cent of the estimated deaths due to malaria occur in sub-Saharan Africa, with 77% of these in children under five years. Although estimated African malaria mortality rates decreased by 54% between 2000 and 2013, much needs to be done to ensure optimization of life-saving malaria interventions. The Pan African Clinical Trials Registry (www.pactr.org; PACTR), launched in 2007, provides online information about completed, ongoing and planned clinical trials in Africa, which can be used to apply the best available strategies to treat and prevent disease, and plan future research. Objectives: To analyse interventions reported in malaria trials conducted in Africa and registered on PACTR between 2007 and 2014, and to compare them against those included in a cohort of pre-2007 trials, to highlight trends and possible gaps in interventions. Methods: We conducted a cross-sectional analysis of trials currently registered on www.pactr.org, and extracted those reporting malaria interventions. Data extraction included number of trials, location, intervention studied, and participant age range. Descriptive analysis was run in MS Excel. Results: Of 388 studies registered on www.pactr.org from 1 May 2007 to December 2014, 51 trials reported on planned or ongoing Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. research on malaria interventions. Of these, 24 reported on treatment, 26 on prevention and one on other interventions. For the prevention trials, seven included drug interventions, 15 reported on vaccines, two on mosquito control, and two on other methods. Compared to the older cohort, there is an increase in the number of vaccines trials (from 15.7% to 57.6%) and a decrease in the number of mosquito control trials (from 26.3% to 7.6%). Other comparative analyses are ongoing. Conclusions: A comparison of African malaria trials pre-2007 and from 2007 to 2014 revealed changes in research trends. Searches of registers such as PACTR are indispensable tools to allow up-to date research monitoring and appropriate planning for future initiatives. P1.094 Does registration will help systematic reviews to be published in journals indexed by Web of Science? A current investigation based on the PROSPERO platform Zhang Y1 , Peng Y1 , Sun X1 , Du L1 , Huang J1 1 The Chinese Cochrane Center, West China Hospital of Sichuan University, China Background and Objective: The PROSPERO platform is an international prospective register platform established by the University of York in 2011. Whether registration will help systematic reviews to be published in journals indexed by Web of Science is not known. The current study aims to investigate published systematic reviews that have been registered on this platform, based on the Web of Science database. Methods: The PROSPERO platform was searched for systematic reviews for which the review status was ‘published’. The last search was performed on 11 January 2015. Two reviewers independently screened the included systematic reviews according to the inclusion and exclusion criteria, and extracted data. The country, journal, time between the registration date and publication date, and funding were all analyzed. Results: A total of 239 systematic reviews were included in the data analysis. The average time between the registration date and published date was almost 14 months. These studies were from 35 countries, amongst which the UK published the most studies (n = 70), followed by Canada and Australia. All these studies were published in 175 journals, of which, PLoS One published 19 studies and ranked first, followed by the Health Technology Assessment and BMJ. With regard to the funding, 130 studies were supported by funders, while 109 studies were not supported by funders. Conclusion: The current evidence indicates that published systematic reviews that have been registered on the PROSPERO platform are relatively few. Most of these studies are from the developed countries. It is still necessary to disseminate the importance of PROSPERO. Cochrane Database Syst Rev Suppl 1–327 (2015) 204 Outcomes/Safety and adverse events P1.095 Endoscopic and open release similarly safe for the treatment of carpal tunnel syndrome. A systematic review and meta-analysis Nikolakopoulou A1 , Vasiliadis H2 , Shrier I3 , Lunn M4 , Brassington R4 , Scholten R5 , Salanti G1 1 Department of Hygiene and Epidemiology, University of Ioannina School of Medicine, Greece; 2 Department of Orthopaedics, University of Ioannina, School of Medicine, Greece; 3 Centre for Clinical Epidemiology, Lady Davis Institute for Medical Research, Jewish General Hospital, McGill University, Canada; 4 Centre for Neuromuscular Disease, National Hospital for Neurology and Neurosurgery, Queen Square, London, UK; 5 Dutch Cochrane Centre and Julius Center for Health Sciences and Primary Care, University Medical Center Utrecht, The Netherlands Background: Analysis of safety data using randomized controlled trials is often challenging because events are typically rare. Various statistical methods have been proposed to model rare events in a frequentist setting. In a Bayesian framework, rare events can be modeled using the binomial likelihood. Evaluation of the comparative safety of endoscopic (ECTR) and open carpal tunnel release (OCTR) poses an additional challenge; the outcomes are often correlated because of bilateral involvement. Moreover, statements about the conclusiveness of evidence are difficult to draw when outcomes are rare. Objectives: To evaluate whether there is evidence supporting a higher risk after ECTR in comparison to the conventional open release using a variety of methods to address rare events and developing a model for dichotomous correlated outcomes. Methods: We included all randomized or quasi-randomized controlled trials in a systematic review and synthesised them in a meta-analysis. Safety was assessed by the incidence of major, minor and total number of complications, recurrences, and re-operations. Different methods to handle rare events and a Bayesian model were compared. A bivariate random-effects model was developed to account for the correlated events. We explored various methods to approximate correlation coefficients based on the number of patients with bilateral involvement. We applied three different methods to decide whether the evidence was conclusive. Results: The assessment of major complications, reoperations and recurrence of symptoms does not favor either of the interventions and all models are consistent in their results. ECTR performs better in terms of minor complications compared to OCTR. Despite the fact that complications were rare, it is unlikely that future evidence will alter these conclusions. Conclusions: Various Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. statistical analyses for correlated and rare events did not suggest a significant benefit of one surgical technique over the other. The strength of evidence in favor of a clinically relevant difference is weak. Further studies are not likely to detect a difference between ECTR and OCTR in terms of safety outcomes. P1.096 Scope and consistency of outcomes reported in randomised trials included in Cochrane systematic reviews of haemodialysis Tong A1 , Williams G1 , Manns B2 , Hemmelgarn B2 , Wheeler D3 , Tugwell P4 , Winkelmayer W5 , Urquhart-Secord R1 , Craig J1 1 TheUniversityofSydney, Australia; 2 UniversityofCalgary, Canada; 3 University College London, United Kingdom; 4 University of Ottawa, Canada; 5 Baylor College of Medicine, USA Background: Inconsistent outcome selection and reporting in clinical trials may limit their capacity to inform evidence-based decision making and the ability to combine findings in systematic reviews. Objectives: To assess the scope and consistency of outcome selection and reporting in trials included in systematic reviews of interventions for patients on chronic haemodialysis. Methods: The Cochrane Database of Systematic Reviews was searched from 1999 to January 2015 for published systematic reviews of interventions for patients on chronic haemodialysis. The description and frequency of outcomes reported within the source randomised controlled trials were assessed. Results: The 18 systematic reviews included 282 trials that reported over 518 outcomes. There was considerable heterogeneity among the outcomes reported. Across all trials, 320 outcomes occurred only once, and 198 outcomes were identified twice. The five most common outcome domains were mortality (50 trials, 18%), blood pressure (46 trials, 16%), calcium (44 trials, 16%), parathyroid hormone (41 trials, 15%) and quality of life (39 trials, 14 %). There was considerable variation in how the domains were measured, the time at which they were measured, and the units and threshold changes that defined an outcome. For example, blood pressure was reported as pre-dialysis, post-dialysis, maximum, difference in pre- and post-dialysis, ambulatory, resting, rise, and time to restore. Conclusions: A wide array of outcomes are reported in trials of interventions in haemodialysis, with large heterogeneity in timing of measurements, units of measure, and threshold changes that define an outcome. This highlights the need for a well-defined set of standardized core outcomes to improve the comparability of trial findings in meta-analysis and to provide greater confidence around treatment decisions for patients undergoing haemodialysis. Cochrane Database Syst Rev Suppl 1–327 (2015) 205 P1.097 P1.098 Quality assessment of oral health-related quality-of-life questionnaires in children and adolescents Inconsistent evaluation of heparin-induced thrombocytopenia in Cochrane systematic reviews: are harm-benefit balance truly being appraised? Zaror C1 , Pardo Y2 , Martinez-Zapata MJ3 , Espinoza G4 , Muñoz-Millán P5 , Ferrer M2 1 CIGES, Faculty of Medicine, Universidad de la Frontera, Chile; 2 Health Services Research Group, IMIM (Hospital del Mar Medical Research Institute), Spain; 3 Iberoamerican Cochrane Centre, Institute of Biomedical Research (IIB Sant Pau), Spain; 4 Department of Public Health, Faculty of Medicine, Universidad de la Frontera, Chile; 5 Department of Pediatric Dentistry and Orthodontic, Faculty of Dentistry, Universidad de la Frontera, Chile Background: Oral health–related quality of life (OHRQoL) is a multidimensional construct that includes a subjective evaluation of the individual’s oral health, functional well-being, expectations and satisfaction with care, and sense of self. The development and validating of specific instruments for children of different ages has allowed better capture of effects related to children’s oral health problems, but only a little is known about the quality of them. Objectives: To perform a quality evaluation of the available evidence on the development process and metric properties of OHRQoL questionnaires in children and adolescents. Methods: To identify all available OHRQoL questionnaires a systematic literature search was performed in MEDLINE, EMBASE, Cochrane, SciELO, and LILACS. The search was complemented through manual review of the references of included articles. Both selection and data extraction were conducted by two researchers independently. Two experts independently assessed all the articles identified for one instrument and applied the Evaluating Measures of Patient Reported Outcomes (EMPRO) tool, which was designed to assess the quality of OHRQoL questionnaires in a standardized way. An overall EMPRO score and six attribute-specific scores were calculated (range, 0 to 100) to describe the quality of instrument performance. Results: The search revealed 2465 articles, of which 321 were duplicates, with 177 full-texts examined after selection based on titles and abstracts. Eleven articles were identified by handsearching. Finally 18 questionnaires and 156 associated studies of their metric properties were identified for evaluation with the EMPRO tool. Eleven were generic instruments to measure QHRQoL, one for patients with hypodontia, one for patients with Down syndrome, one for patients with fixed appliances, two for discomfort and two evaluate impact of OHRQoL on parents and families. Conclusions: Knowing the quality of the instruments used to assess the OHRQoL allow us to apply the most appropriate for clinical practice and research. We plan to present the complete results and definitive conclusions at the Cochrane Colloquium 2015. Junqueira D1 , Faria JC1 , Fontoura M1 , Menezes de Pádua CA1 1 Universidade Federal de Minas Gerais, Brazil Background: Bleeding is the main recognised adverse effect of heparins. However, other adverse effects, such as heparin-induced thrombocytopenia (HIT) can also be expected. HIT is an immune reaction, frequent in some clinical scenarios, and associated with paradoxical increase in the risk of thromboembolic complications. Objectives: To assess the consistency of the assessment of HIT in Cochrane Systematic Reviews (SRs) evaluating heparin interventions for prophylaxis or treatment of thromboembolism. Methods: The Cochrane Database of Systematic Reviews was searched from inception to March 2015 for SRs including any heparin (unfractionated heparin (UFH), or low molecular weight heparin (LMWH)) for prevention or treatment of thromboembolism. Detailed information obtained from the SRs included: (1) statement of thrombocytopenia or HIT as outcome (primary or secondary); (2) statement of a precise definition for HIT (HIT is defined as a relative reduction in platelet count of about 50% (even if the platelet count at its lowest remains > 150 x 109/L) occurring within five to 14 days after the start of heparin therapy and confirmed by antibody tests); (3) reporting of HIT in the Results section qualitatively or quantitatively. Results: Fifteen SRs published from 2002 to 2015 were identified. The reviews investigated a wide range of clinical scenarios and most were assessed as up-to-date after 2010 (66.6%). Six reviews (40%) specified thrombocytopenia as a secondary outcome and three (20%) cited HIT as secondary outcome. Only one review provided a clear, though incomplete, definition of HIT. In the Results section, one SR reported no cases of HIT without defining the reaction, one presenting a definition for HIT and reported that participants from the included trials were not investigated for HIT, and another was an empty review. Conclusions: HIT is a harmful effect of heparins not evaluated in Cochrane Reviews. The assessment of the harm-benefit balance of an intervention cannot be limited to the main recognised adverse effect. Improvements are needed in order to reduce bias in the assessment of HIT, and probably other adverse effects, in Cochrane Reviews. P1.099 Estimating the occurrence of non-standardised adverse outcomes: the case of lipodystrophy Junqueira D1 , Lana LG1 , Perini E1 , Menezes de Pádua CA1 1 Universidade Federal de Minas Gerais, Brazil Background: Retrieving appropriate adverse drug reaction data from epidemiological studies is a challenging task, even Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Cochrane Database Syst Rev Suppl 1–327 (2015) 206 when well-established outcomes (e.g. HIT- heparin-induced thrombocytopenia) are taken into account. Lipodystrophy (LD) is a disfiguring adverse reaction to antiretroviral therapy (ART), which has been reported without consensus on the case definition. Objectives: To estimate the frequency of LD among HIV-infected patients on ART in order to establish occurrence estimates according to diagnostic methods and case definitions reported in the literature. Methods: A systematic review of non-randomised studies (NRS) was carried out through a sensitive search of MEDLINE, CINAHL, LILACS, EMBASE and International Pharmaceutical Abstracts databases - up to June 2014. NRS reporting LD as a primary or secondary outcome and comparing HIV-infected patients on different ART for at least six months were considered eligible. The main measure of occurrence was the incidence and prevalence of LD. Results: Twenty studies (12 cross-sectional and eight prospective) were included in the systematic review. Most studies (80%) provided detailed information on the definition of LD, describing the anatomical sites, and four (20%) reported the outcome in vague and unspecific terms. All the studies investigated morphological alterations in the abdomen, face and limbs, and other anatomical sites. The diagnostic methods varied between studies (n = 4; 20%), with the concomitant evaluation of the physician and the patient being the most used (n = 9; 45%). Considering the reported nomenclature, the prevalence of LD ranged between 11.7% and 67.8%. The incidence of LD ranged between 8.1% and 52.0%. Conclusions: The very heterogeneous case definitions used in the NRS precluded the establishment of point estimates of LD. Instead, we were able to provide rough interval estimates. This might be a first effort to demonstrate that: (i) improvement in the reporting of LD is needed, and (ii) systematic reviewers must be aware of the challenges in estimating adverse drug reactions that require better clinical definition and depend on diagnostic methods to be identified. P1.100 Reporting health related quality of life in critical care RCTs Roth D1 , Tscherny K1 , Heidinger B1 , Havel C1 , Arrich J1 , Gamper G2 , Herkner H1 1 Medical University of Vienna, Department of Emergency Medicine, Austria; 2 Universitätsklinikum St. Pölten, Austria Background: Health-related quality of life (HRQoL), defined as ‘the functional effect of an illness and its treatment upon a patient as perceived by the patient’, might provide better information about the actual impact of a therapy; detect less obvious or unexpected effects; and is among the most relevant outcomes. A Cochrane Methods Group focusing on patient-reported outcome measures has been formed in 1999, and analysis of the reporting of quality of life has been recommended for all Cochrane Reviews since 2003. The frequency of reporting HRQoL in randomized controlled trials (RCTs) on critical care in general has however not been yet evaluated. Objectives: To assess how many Cochrane Reviews on critical care included HRQoL as an outcome, Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. evaluate the frequency of reporting of HRQoL in RCTs in critical care medicine, and to analyze the differences in study characteristics between those reporting HRQoL and those not. Methods: To access a cohort of studies, the Cochrane Database of Systematic Reviews was searched for reviews on critical care published after 2000. All reviews in the RCTs contained therein were included in our analysis. Study characteristics such as sample size, type of intervention, disease spectrum, intensive care unit (ICU) setting, hospital setting, funding, effect size, number and type of outcomes, risk of bias, were extracted. HRQoL-specific data included whether HRQoL was reported, and which instrument was used. Results: The reference lists of 136 reviews, 53 (39%) of which mentioned HRQoL as an outcome, were reviewed. A total of 468 individual studies representing 88,585 patients were included; 39 (8%) reported HRQoL as an outcome, one (0.2%) study as a primary outcome. Most commonly used scores were the Glasgow Outcome Scale (8; 21%), Modified Rankin Score (6; 15%), Rankin Score (4; 10%) and Barthel Index (3; 8%). No differences were found between studies reporting HRQoL and those not. Conclusions: Although measuring HRQoL is recommended for all Cochrane Reviews, only 39% defined it as an outcome. Reporting HRQoL is even less common in individual studies of critical care, with only 8% of studies including measures of HRQoL. P1.101 Psychometric evidence for outcome measures used in patients undergoing total hip arthroplasty: a systematic review Gagnier J1 , Huang W1 , Mullins M1 1 University of Michigan, USA Introduction: The number of clinical trials testing variations in total hip arthroplasty (THA) are increasing. Research suggests that these studies use a heterogeneous array of outcome measures and that many of these measures have questionable reliability and validity. This creates difficulty when attempting to perform systematic reviews of these trials and in using the results of clinical trials to inform decision making. Objectives: Our objectives were to: (1) identify currently available patient reported outcome (PRO) measures used in THA patients; (2) appraise the methodological quality of the studies that evaluate the identified instruments; (3) assess the psychometric evidence of these instruments; and (4) provide a summary of the overall evidence for and against each included questionnaire. Methods: We searched MEDLINE, EMBASE, CINAHL, SportDiscus, the Cochrane Library, Scopus (1931 to 2014) and reviewed reference lists and contacted experts to identify relevant articles. We included articles on the development or evaluation of a psychometric property of a PRO measure in those undergoing THA and published in English. The study risk of bias was appraised using the ‘consensus-based standards for the selection of health status measurement instruments’ checklist and psychometric evidence assessed using criteria proposed by Terwee et al. Cochrane Database Syst Rev Suppl 1–327 (2015) 207 Finally, we combined these assessments to arrive at a synthesis of overall evidence for each PRO measure. Results: We included 49 unique studies testing properties of 18 different PRO measures. The Western Ontario and McMaster Universities Osteoarthritis (WOMAC) Index, the Oxford Hip Score (OHS), and the Harris Hip Score had the most studies assessing them. The OsteoArthritis Knee and Hip Quality Of Life (OAKHQOL), OHS, Patient Specific Index (PASI), and WOMAC had the highest overall evidence. Most included studies were poorly reported and very few psychometric properties were tested in the literature. Conclusion: We found that very few instruments had sufficient psychometric properties. We recommend that reporting and methodologic guidelines be developed for the assessment of psychometric properties of PRO measures. P1.102 Psychometric evidence for outcome measures used in patients undergoing total knee arthroplasty: a systematic review Gagnier J1 , Huang W1 , Mullins M1 1 University of Michigan, USA Introduction: Total knee arthroplasty (TKA) is a viable treatment option for improving pain and function due to severe osteoarthritis. Current recommendations suggest using patient reported outcome (PRO) data to assess TKA outcomes. Many PRO instruments have been developed and compared, but this research often has questionable quality and does not adhere to methodologic and reporting guidance. Objectives: (1) to identify currently available PRO questionnaires used in TKA patients; (2) to assess the risk of bias of studies that evaluate psychometric properties of identified instruments; (3) to assess the psychometric evidence of these instruments; and (4) to provide a summary of the overall evidence across each PRO measure. Methods: We searched MEDLINE, EMBASE, CINAHL, SportDiscus, the Cochrane Library, Scopus (1931 to 2014), reviewed reference lists and contacted experts to identify relevant articles. We included articles on the development or evaluation of a psychometric property of a PRO measure in those undergoing TKA and published in English. The study risk of bias was appraised using the ‘consensus-based standards for the selection of health status measurement instruments’ checklist and psychometric evidence assessed using criteria proposed by Terwee et al. Finally, we combined these assessments to arrive at a synthesis of overall evidence for each PRO measure. Results: We included a total of 59 studies on 15 different PRO instruments, with Western Ontario and McMaster Universities Osteoarthritis (WOMAC) index and the Oxford Knee Score (OKS) being the most frequent. Most instruments had poor evidence, and only the OKS and the Osteoarthritis Knee and Hip Quality of Life questionnaire (OAKHQOL) had any excellent evidence. OAKHQOL had the highest overall performance with three properties having positive evidence. Conclusions: Despite many studies on Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. PRO instruments for TKA, few are of good quality. As a whole, additional studies are needed throughout the literature to produce quality evaluations of PRO instruments for TKA. We recommend that reporting and methodologic guidelines be developed for the assessment of psychometric properties of PRO measures. P1.103 The higher quality, the more convincing: a cross-sectional study of randomized controlled human trials on drug safety Zhang L1 , Chen Q2 , Zhao P1 , Jia P1 , Zhang M1 1 Chinese Evidence-Based Medicine Centre, West China Hospital, Sichuan University, China; 2 West China School of Public Health, Sichuan University, China Background: Can drug safety analyses in randomized controlled trials answer drug safety issues accurately? Regrettably, for much research the answer is no. Findings from drug safety analyses may be misleading, potentially resulting in clinical threat and health decision making. Few studies have investigated the reporting and conduct of drug safety analyses and a number of important questions remain unanswered. Objectives: To describe the methodological quality of drug safety research in randomized controlled trials with a modified version of the Cochrane ‘Risk of bias’ (RoB) tool. Methods: We conducted a cross-sectional study of 165 eligible randomized controlled human trials on drug safety, which we randomly selected from MEDLINE, Embase and the Cochrane Central Register of Controlled Trials (CENTRAL) in 2014. Teams of two reviewers independently screened full texts of reports for eligibility, and abstracted data, using standardized, pilot-tested extraction forms. We assessed the risk of bias in these studies According to a modified version of the Cochrane RoB tool (we considered random sequence generation; allocation concealment; blinding of participants, caregivers, and outcome assessors; adjudication of adverse events; prognostic balance between treatment groups; and incomplete outcome data). We conducted a statistical analysis to describe current situation about the methodological quality of drug safety research in randomized controlled trials. Results: Within these studies, only 20% specifically mentioned the randomized method, but 77% mentioned randomization without specifying methods. Similar results appeared for allocation concealment; patients and caregivers were blinded in 48%, but outcome assessors in only 24%. There were few studies that made judgements of adverse events. Conclusions: Currently, qualities of randomized controlled trials about drug safety research are uneven. Research of high quality with sufficient reports can get more approvals. Our findings will contribute to a set of recommendations to optimize the conduct and reporting of drug safety analysis. Cochrane Database Syst Rev Suppl 1–327 (2015) 208 P1.104 When you conduct a systematic review about drug safety, you will run into these issues Zhang L1 , Chen Q2 , Zhao P1 , Jia P1 , Zhang M1 1 Chinese Evidence-Based Medicine Centre, West China Hospital, Sichuan University, China; 2 Sichuan University West China School of Public Health, China Background: Safety analyses in randomized trials examine whether adverse events were caused by treatment. However, findings from safety analyses may be misleading, potentially resulting in suboptimal clinical and health decision-making. Few studies have investigated the reporting and conduct of safety analyses and a number of important questions remain unanswered. Objectives: To assess study characteristics associated with the results (such as the quality and credibility of research results) of safety analyses. Methods: We conducted a systematic review of 165 eligible, randomized, controlled human trials on drug safety, which we randomly selected in MEDLINE, Embase and the Cochrane Central Register of Controlled Trials (CENTRAL) until the number of them met our required sample size. Teams of two reviewers independently screened full texts of reports for eligibility, and abstract data, using standardized, pilot-tested extraction forms. We conducted univariable and multivariable logistic regression analyses to examine the association of pre-specified study characteristics (did it make an accurate definition of safety prior; did it describe the frequency clearly (i.e. times or person-times); did it accurately define denominator of adverse events) with the effect of drug safety analyses. Progress and Discussion: Currently, we are extracting the data of these publications. Although reports of controlled trials are recorded sufficiently with regard to curative effect, they are not reported to the same level for safety. A clear understanding of safety analyses, as currently conducted and reported in published randomized controlled trials, will reveal both strengths and weaknesses of this practice. Our findings will contribute to a set of recommendations to optimize the conduct and reporting of safety analyses. P1.105 Can assessors in a trial of psychotherapy be successfully blinded? Watanabe N1 Japanese Cochrane Branch, Japan 1 Background: Blinding is a key element of treatment evaluation in systematic reviews, and is considered more difficult to obtain in trials assessing psychosocial treatment than in those on pharmacotherapy. Objectives: To evaluate the concordance rate and kappa values for agreement between the right allocation and the guessed allocation by blinded assessors in a randomized controlled trial (RCT) Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. on psychotherapy. Factors associated with concordance between actual assignments and guessed assignments are also explored. Methods: Data on both actual and guessed assignments by six blind assessors were obtained at fourand eight-week assessments in an RCT assessing the added value of brief behavioral therapy for insomnia over treatment as usual for residual depression and refractory insomnia in 37 adults at three psychiatric outpatient departments. The severity of depression was assessed by blind raters. Neither the patients nor physicians of TAU were blind to allocation. However, all patients were requested not to reveal their allocated treatment to the assessors for the severity of depression. After each assessment, an assessor guessed which group the patient had been assigned to, making it possible to examine whether the blinding was successful. Information about the degree of confidence in and reasons for guessing such was also gathered from the assessors. Results: The concordance rates and kappa values for agreement between the actual allocation and the allocation guessed by blind assessor at each assessment were 56.7% and 0.15 at four weeks and 70.2% and 0.41 at eight weeks, respectively. This indicated that the blinding of the assessors was satisfactory. With regard to factors associated with the right allocation, only ‘intuition’ of the assessors was statistically significant (P value = 0.02). Conclusions: Assessors in a trial on psychotherapy can be successfully blinded, where a sufficient effort is made. Reference Watanabe N, Furukawa TA, Shimodera S, et al. Psychother Psychosom 2013; 82:401–403. P1.107 Difficulty in combining results of primary studies in the addiction field: need to enhance consistency of reported outcomes Minozzi S1 , Mitrova Z1 , Saulle R1 , De Crescenzo F2 , Amato L1 1 Cochrane Drugs and Alcohol Group, Italy; 2 Institute of Psychiatry and Psychology Catholic University of the Sacred Heart, Rome, Italy Background: The COMET initiative outlined the need for the development and application of standardized sets of outcomes agreed by all stakeholders, including patients, known as a ‘core outcome set’. In order to allow the results of trials to be compared, contrasted and combined as appropriate, the defined core outcome set should be measured and reported in all trials of a specific health area. This need is particularly relevant in the field of addiction, where there is a considerable heterogeneity in outcomes and measures used. Objectives: To analyze the variety of outcomes and measures used in randomized controlled trials assessing the effectiveness of pharmacological intervention for psychostimulant misuse, taken as an example to identify the most used outcomes and measures. Methods: Cochrane Systematic Reviews published by the Cochrane Drugs and Alcohol Group up to the end of June 2015 and assessing Cochrane Database Syst Rev Suppl 1–327 (2015) 209 the effectiveness of psychosocial interventions for drug and alcohol misuse will be analyzed. The following data will be reported: number of reviews and of included studies, a complete list of clinical outcomes and of different ways to measure each outcome, outcomes for which pooling of studies was possible and outcomes reported in the ‘Summary of findings’ table(s). Results: Our preliminary results based on nine reviews with 121 included trials showed that the highest heterogeneity was found for drug use outcomes: use of dichotomous versus continuous outcomes; point abstinence versus continuous abstinence; different length of abstinence period considered; different time points for outcome assessment; self- reported versus objectives measures. Conclusions will be presented at the Cochrane Colloquium 2015. Conclusions: Our findings could be used in the process of defining an agreed core outcome set Consumer issues and shared decision-making P1.110 Synthesis of health system barriers and facilitators to scaling up vaccination communication interventions in Cameroon, Mozambique and Nigeria Muloliwa AM1 , Cartier Y2 , Ames H3 , Oku A4 , Bosch-Capblanch X5 , Cliff J6 , Glenton C3 , Hill S7 , Kaufman J7 , Oyo-Ita A4 , Rada G8 , Lewin S9 1 Provincial Directorate of Health, Nampula, Mozambique, Mozambique; 2 International Union for Health Promotion and Education, France; 3 The Norwegian Knowledge Centre for the Health Services, Norway; 4 University of Calabar, Nigeria; 5 Swiss Tropical and Public Health Institute, Switzerland; 6 Universidade Eduardo Mondlane, Mozambique; 7 La Trobe University, Australia; 8 Catholic University of Chile, Chile; 9 The Norwegian Knowledge Centre for the Health Services in Norway and Health Systems Research Unit, South African Medical Research Council, South Africa Background: Communication between healthcare providers and individuals and communities is a key factor affecting childhood vaccination uptake worldwide. Effective communication is critical to successful, sustainable vaccination programmes. However, there is limited evidence specific to low income settings (LISs) to guide programme managers in implementing communication interventions. The ‘Communicate to vaccinate’ (COMMVAC 2) project aims to build research evidence on the effects of communication with parents and communities to improve childhood vaccination uptake in LISs. Objectives: To synthesise evidence on health system barriers and facilitators to scaling up vaccination communication interventions in three countries where the COMMVAC 2 project is being implemented: Cameroon, Mozambique and Nigeria. Methods: We will identify eligible documents through searches of relevant electronic databases; consultation with the COMMVAC 2 partners, experts and key Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. agencies, such as the World Health Organization; and through internet searching. We will include documents that focus on factors affecting the scale-up of childhood vaccination communication in the study countries. Data from these documents will be combined with data from key informant interviews in each country. Framework synthesis will be used to understand how health system factors affect the implementation of vaccination communication interventions at scale. Results: Data collection in countries and consultation with experts is underway. Database searching has been completed and screening is underway on 146 full text documents covering Nigeria (n = 98), Cameroon (n = 20), Mozambique (n = 15) and a range of countries (n = 13). Preliminary results will be presented during the Cochrane Colloquium. Conclusions: Understanding the factors associated with the effective implementation at scale of communication interventions for childhood vaccination will inform the selection and adaptation of interventions to improve vaccination coverage. P1.111 How evidence-based are Print- and Online Mass-Media in Austria? A quantitative analysis Kerschner B1 , Wipplinger J1 , Klerings I1 , Gartlehner G1 1 Cochrane Austria, Danube University Krems, Austria Background: Print and Online-Media are important sources of health information for lay people. Previous studies suggest that mass media often exaggerate medical facts, however so far, no quantitative data on the degree of exaggeration exist to our knowledge. Objectives: To quantify the degree of deviation of medical news stories from the actual evidence supporting these stories. Methods: Between May 2011 and June 2014, the media watch blog Medizin-Transparent.at covered 219 health topics that had been the subject of mass media reports in Austria. We searched the WISO newspaper database and Google for all Austrian print and online news stories about these topics published within a time window of three months of the original news stories. The strength of evidence for these topics was determined in rapid review processes and rated on a modified GRADE scale and was subsequently compared to subjective ‘implied evidence’ in newspapers. The latter was rated on a self-developed scale on the basis of the modified GRADE scale by two raters independently. Results: Our search yielded 990 media articles in Austrian print and online media covering the 219 prespecified health topics. In 59% of these articles, the evidence for medical facts was reported in a strongly distorted way (exaggerated or understated), only 11% adequately reflected the real strength of evidence. While 61% of articles implied the highest level of evidence for the reported effects, the real evidence was at the highest level for only 3% of articles. Compared to quality media, tabloid media report in a more distorted way. However, this is mainly due to a different choice of covered subjects, since the degree of distorted reporting Cochrane Database Syst Rev Suppl 1–327 (2015) 210 did not differ significantly between the tabloid and quality media for subjects covered by both. Conclusions: More than half of Austrian newspaper and online media reports strongly exaggerated or understated medical facts when compared to the actual evidence for these facts. Tabloid media reported in a slightly more distorted way than quality media. Our results confirm previous reports about serious incorrect reporting of medical facts in mass media. P1.112 Exploring the understanding of consumers and clinicians about network meta-analysis results and their usefulness to make decisions: an Italian survey Del Giovane C1 , Tramacere I2 , Filippini G2 , D’Amico R1 1 Italian Cochrane Centre, Italy; 2 Fondazione I.R.C.C.S. Istituto Neurologico Carlo Besta, Italy Background: Network meta-analysis (NMA) provides a summary of the best evidence already available in the literature and can be extremely useful for different types of stakeholders such as consumers and clinicians. Therefore, it is important that this method is fully understood and properly recognized, especially in multiple sclerosis (MS). Objectives: To investigate: (1) to what extent consumers and clinicians are aware of NMAs; (2) how they understand and interpret the outputs of NMAs; and (3) how the way the results of NMAs are summarized and presented could be enhanced. Methods: The assessment was carried out through a survey. We focused on a specific clinical condition, MS, and the sample of the two types of stakeholders were patients with MS and neurologists. A web-based questionnaire was developed for each group. The questionnaires included questions that intend to: explore the general knowledge and understanding of the risks and benefits of health interventions from randomized controlled trials, systematic reviews and meta-analysis, and NMA; investigate the understanding, usefulness and how to enhance graphical representations of results from a NMA on MS. Text and graphs examples were shown in the questionnaires. Results: The results will be available and presented at the Cochrane Colloquium as the survey will be launched in April 2015. Conclusions: We expect to evaluate and improve the way the results are presented and understood by patients who need to be well informed and receive the best available therapy, and clinicians who need to decide which treatment to use for a specific disease. P1.113 What German hospital pharmacists want to know: cross-sectional survey on information behaviour and information needs Bollig C1 , Günther J2 , Suter K3 , Hoppe-Tichy T4 , Antes G1 , Lang B1 1 Cochrane Germany, Germany; 2 PharmaFacts GmbH, Freiburg, Germany; 3 Hospital-Pharmacy, University Hospital Basel, Switzerland; 4 Pharmacy Department, University Hospital of Heidelberg, Germany Background: The main role of hospital pharmacists in Germany is to prepare and dispense drugs, but they are also essential contact people for physicians, nurses and patients for questions related to pharmacotherapy. To provide high-quality information, hospital pharmacists need access to unbiased and up-to-date drug information sources. Objectives: Investigation of information habits of hospital pharmacists and identification of barriers limiting appropriate information retrieval in daily routine and continuous education. Methods: Since October 2013 Cochrane Germany at the Medical Center-University of Freiburg and the Chair of Scientific Journalism at the Technical University (TU) in Dortmund (Germany) have investigated questions relating to medical knowledge translation. A subproject at Cochrane Germany analyses the information-seeking behaviour and the information provision for German hospital pharmacists. Therefore, approximately 1700 members of the German Society of Hospital Pharmacists will be invited by e-mail to complete a voluntary and anonymous online survey. The questionnaire consists of open-ended and closed questions about the importance of different categories of information and the frequency of drug information questions in their daily routine, the most commonly consulted sources, the amount of time available and necessary, their desires concerning access to drug information sources, and demographics about their education and practice setting. Results: Results of the survey will be available and analyzed in spring 2015 and will be presented at the Cochrane Colloquium 2015. Conclusions: Our questionnaire will identify whether language barriers, lack of time or limited access are obstacles to hospital pharmacists using drug information sources. Moreover, we will investigate if the information sources used fulfil the requirements of hospital pharmacists. P1.114 HTA guide Pertl D1 1 Gesundheit Österreich GmbH, Austria Background: Gesundheit Österreich GmbH (GÖG), on behalf of the Austrian Federal Ministry of Health, has drawn up an internet guide on health technology assessment (HTA). HTA goes back to the late 1970s, when political decision makers voiced demands for well-founded, comprehensive information on health technologies, possible courses of Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Cochrane Database Syst Rev Suppl 1–327 (2015) 211 action and their consequences. It is generally recognised that a central aspect of HTA is to provide a scientific evaluation of health technologies for the purpose of political consultancy. On the basis of the best available evidence, health technology is objectively assessed in a transparent and systematic process, taking into account the specific context in which it will be applied, in order to provide a basis for decision-making by political actors. What is common to all definitions is that HTA goes beyond assessing the mere effectiveness of health technology. Since the beginning of HTA in the late 1970s, various types of efforts have been made to force HTA for decision-making. Objectives: The objective of the HTA guide is to provide quick and easy access to national and international HTA information for the general public and decision makers. In addition, the HTA guide aims to pool links of relevant institutions in the field of HTA. Methods: The HTA guide is revised and updated regularly, the information is available in English and German. New contents are searched for via the internet. The identification of new contents is also facilitated by links of the membership of GÖG to several expert networks. In addition, interested institutions or persons can contact the webmaster with regard to the content or with technical questions. Results: The HTA guide gives an overview of definitions of HTA, handbooks, guides and glossaries, lists institutions, outcomes and sources of HTA in Austria, in selected other countries and across countries. It also provides links to HTA newsletters, education and training offers, and free and for-pay databases. The HTA guide is available under http://hta-guide.biqg.at/. P1.115 Building partnerships with the public by learning about Cochrane evidence Nunn J1 , Synnot A1 , McDonald S2 , Allen K2 , Hill S1 1 Cochrane Consumers and Communication Review Group, Australia; 2 Cochrane Australia, Australia Background: One of the goals of Cochrane’s Strategy 2020 is to make Cochrane the ‘home of evidence’ and to build greater recognition of our work amongst the public, including health practitioners, consumers, policymakers and guideline developers. In order to develop stronger partnerships with the public, Cochrane Australia and the Consumers and Communication Review Group are working together to deliver a series of learning and development events for Australian consumers. Objectives: To increase public understanding of health research and the role of evidence in informing health practice and policy. Methods: Building on successful work in the UK, we are developing learning resources and activities to support the public to understand and value evidence from systematic reviews. We will take a partnership approach, working with local health service providers to create free learning and development opportunities for the public, patients, consumers and volunteers. Specific learning objectives include participants being able to explain the value of systematic reviews, how this knowledge can be applied Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. and how they can support the work of Cochrane. The events will use a series of learning resources, facilitated discussions and activities in order to achieve specific learning outcomes. Results: The learning and development opportunities will be delivered in mid-2015. We have completed an initial scoping exercise and identified potential partner organisations, who are helping us shape both the content and the delivery model. We will be able to report on the success and impact of the project against planned objectives and measurable outcomes. Conclusions: This work has implications for the way in which Cochrane can support the public to understand and value the evidence from Cochrane Reviews. We hope to build on this work and create a sustainable and scalable model, working in partnership with relevant organisations. P1.116 Discussing the risks and benefits of opioid use for chronic non-cancer pain (CNCP) Harris J1 American College of Occupational and Environmental Medicine; The Permanente Medical Group, USA 1 Background: The use of opioids for chronic non-cancer pain (CNCP) has increased rapidly in the last 15 years. However, the effectiveness of opioids for CNCP remains unsupported by good quality evidence, and adverse effects on multiple organ systems are increasingly documented. Opioids are often used at high doses and/or in combination with other psychoactive medications. CNCP has been regarded as purely physical, although its etiology is unknown in most cases, and it is often associated with psychiatric conditions and emotional distress. Objectives: To create a knowledge base and framework for informed discussion of the risks and benefits of opioid and other medication use for CNCP. Methods: A community-wide coalition of public health, medical, education, law enforcement and treatment professionals in Marin County, California (Rx Safe Marin) developed evidence-based provider recommendations for this discussion, and associated community education resources. A parallel effort is underway at medical groups in the County. The American College of Occupational and Environmental Medicine has included recommendations for this discussion in its recently updated opioid guideline as well. Results: The evidence for the effectiveness of opioids for CNCP and adverse effects were reviewed. A framework for periodic informed consent and shared decision making discussions was developed, that was similar in all three organizations. Provider material includes a series of case studies and video vignettes for communication about difficult issues and patient concerns. Rx Safe Marin and associated provider groups are monitoring adverse event rates and prescription rates, and working on documentation of shared decision making and informed consent. Conclusions: The wide spread effects of opioids on multiple organ systems, beyond addiction, dependence, overdose and death, balanced against the still unproven effectiveness of opioids for CNCP should be carefully Cochrane Database Syst Rev Suppl 1–327 (2015) 212 considered in a shared decision-making/informed consent framework when making treatment decisions. P1.117 How willing are the public to pay for anti-hypertensive drugs for primary prevention of cardiovascular disease: a survey in a Chinese city Tang J1 , Wang W2 , An J3 , Hu Y4 , Cheng S3 , Griffiths S2 1 TheHongKongBranchoftheChineseCochraneCentre, China; 2 The Chinese University of Hong Kong, China; 3 Baotou Medical College, China; 4 Peking University Center for Evidence-Based Medicine and Clinical Research, China Background: Current recommendations on drug treatment of hypertension for primary prevention of cardiovascular disease are primarily determined by the evidence of effectiveness, disregard the resources available and values of people, and recommend a universally fixed-risk cutoff for initiating drug treatment. The guidelines may have over-estimated the willingness of the public to accept and pay for these drugs and a fixed cutoff may not fit all populations. Moreover, the public may have been misinformed and are unable to make the right decision even if they are consulted. Objective: To describe the gap between current policy and what the public truly want. Methods: A cross-sectional survey with face-to-face interviews of rural and urban residents in northern China. Results: A total of 887 rural residents and 921 urban residents were interviewed with a response rate of 97%. Of these, 95% (95% confidence interval (CI) 94% to 96%) said they would take anti-hypertensive drugs if they had hypertension, although 91% (95% CI 89% to 92%) said they did not have sufficient knowledge to make a decision; 78% (95% CI 76% to 80%) believed that anti-hypertensive drugs were primarily to lower blood pressure or relieve symptoms. They over-estimated the cardiovascular risk of untreated hypertension by approximately 12 times and the absolute benefit of drug treatment by 20 times. Given the actual absolute benefit of the drugs, only 23% (95% CI 21% to 25%) were willing to pay the current annual cost of RMB 500 (USD 73.30) for these drugs. Given the current cost, they were, on average, willing to pay for the drugs only when the five-year cardiovascular disease risk was as high as 35% (95% CI 31% to 38%) or even higher. Conclusion: The public in China are significantly misinformed and considerably over estimate the risk of hypertension and the benefit of treatment. The public’s willingness to pay for anti-hypertensive drugs is much lower than the current guidelines implicitly assume. The willingness to pay should be considered, along with other factors, when prescribing anti-hypertensive drugs to an individual patient or making hypertension guidelines for a population. P1.118 Cochrane initiatives in Nepal: reflections on progress in improving evidence-informed decision-making Aryal K1 , Dhimal M1 , Stewart M2 , Dhakal P1 , Mehata S3 , Rana S4 , Tharyan P5 , Karki K1 1 Nepal Health Research Council, Nepal; 2 Ministry of Health and Population, Nepal; 3 Nepal Health Sector Support Programme, Nepal; 4 London Borough of Merton, United Kingdom; 5 Cochrane South Asia, India Background and Objectives: In 2014, the Nepal Health Research Council (NHRC) and Cochrane South Asia began collaborating formally in order to foster an evidence-enabling environment within the Nepali health sector. This poster reflects on progress and outlines future directions. Methods: Critical reflection that incorporates the perspectives of representatives from government, research institutions, and non-governmental organizations (NGOs) in Nepal. Results: The crucial impetus for collaboration was a workshop attended by key NGO and governmental stakeholders, including the Secretary for Health, at which the NHRC was unanimously designated as the nodal agency for Cochrane-related activities. Subsequent activities included workshops for policy-makers and planners, training on Cochrane methods for researchers, and identification of research priority areas. The NHRC was subsequently invited to provide technical support in developing the Third Nepal Health Sector Plan (NHSP III), which outlines priorities for 2015 to 2020. As a result, the recently-developed NHSP III is vastly more evidence-informed than the previous two plans. Conclusions: We attribute our successes to date to: engaging policy makers early; finding an effective nodal agency that was acceptable to stakeholders; developing and leveraging formal and informal networks; facilitating the participation of key stakeholders at the Hyderabad Colloquium; and establishing evidence-support teams that seek out, synthesize, and present contextualized evidence in digestible formats such as policy briefs, thereby gaining buy-in from high-level stakeholders. Through 2015 we plan to continue engaging government officials and to provide training for in-country research teams. We also hope to establish a branch of Cochrane South Asia in Nepal by 2016. P1.120 Clinical decision support system: an effective pathway to promote shared decision making between clinician and patient. Jia P1 , Zhang L1 , Zhao P1 , Zhang M1 1 Chinese Cochrane Center, China Background: A clinical decision support system (CDSS) has been defined as a computerized system that uses case-based Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Cochrane Database Syst Rev Suppl 1–327 (2015) 213 reasoning to assist clinicians in assessing disease status, in making a diagnosis, in selecting appropriate therapy or in making other clinical decisions. Related studies showed that CDSS can promote shared decision making and improve clinical decisions. However, little is known about the ways and factors that are most effective in promoting shared decision making between clinicians and patients. Objectives: To provide a rigorous, evidence-based assessment of the ways and factors that are most effective in influencing shared decision making between clinicians and patients. Methods: We searched the MEDLINE, EMBASE, Cochrane Library and four main Chinese databases including CBM, VIP, CNKI and Wanfang using the search term ‘decision support systems, clinical’, ‘decision-making, computer assisted’, ‘CDSS’, ‘shared decision making’, ‘shared determinant’ etc from 1977 to March 2015. Microsoft Excel 2007 was used to perform data extraction and analysis. Results: We included a total of 340 studies in our study that ranged from 1997 to 2015. The number of studies reached its peak for 2010, with 56 studies accounting for 17.47% of the years’ publication. The CDSS can promote shared decision making in five major ways: timely alerting (37.5%), transparency interpreting (17.4%),scientific critiquing (8.6%), reasonable assisting (8.0%), diagnosing and managing (6%).Several practical factors contribute to the success of CDSS in promoting shared decision making: (1) automatic provision of decision support; (2) an intuitive user interface; (3) timely decision support; (4) providing actionable recommendations that are succinct and relevant to patient care; and (5) knowledge skill of patient and clinician. Conclusions: There are a great number of ways and factors that influence the use of CDSS for promoting shared decision making between clinicians and patients. Further research needs to be conducted before there is widespread dissemination into clinical practice. P1.121 Patient participation in health care in China: what ways promote patient participation and engagement? Jia P1 , Zhang L1 , Zhao P1 , Zhang M1 1 Chinese Cochrane Center, China Background: Patient participant in health care has been a concern since the complex social movements of the 1960s. The doctor-patient dispute has been increasingly prominent in recent years in China. Related studies showed that the communication openness was seriously jeopardised by the lack of trust between health workers and patients. Objectives: To investigate the ways of promoting the participation of the patient in health care in China. Methods: We searched four Chinese databases including CBM, VIP, CNKI and Wanfang using the search term ‘patient participant’, ‘patient involvement’ ‘health care’ and ‘clinical care’ from 2000 to February 2015. Results: Eighty relevant studies were published from 2000 to 2015 but only 22 meet our criteria. The Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. number of studies relevant to patient participant in health care was increasing from the year 2010 to 2014, with two, three, five and 10 respectively. Five categories of ways were identified that promote patient participant in health care: education by clinicians (8), self-monitoring (5), communicating information (3), family’s participation (2) and education by internet (2). The participants were diabetics, children, women with cesareans and patients of plastic surgery. Among the 20 studies, the institutions of the first authors mostly were universities (65%), hospitals (11%), and colleges (9%), respectively. Conclusions: Patient participant in health care is still in its infancy in China. Although participation in health care has been emphasized for many years, there is not much evidence to support this. Hopefully ‘participation in health care’ will be a focus in future in China. P1.122 Patient safety education: a cross-sectional survey of knowledge, skill and attitude of medical students on patient safety in China Jia P1 , Zhao P1 , Zhang L1 , Zhang M1 1 Chinese Cochrane Center, China Background: Reducing medical errors and improving patient safety have become the priority of health care around the world. China has been carrying out patient safety education in medical students for the past 10 years. Medical students should be able to recognize unsafe conditions, and systematically report errors and near misses. Objectives: To assess the knowledge, skill and attitude of medical students on patient safety. Methods: This is a descriptive and cross-section study. A questionnaire of 31 items was developed based on the World Health Organization (WHO) study for patient safety curriculum guide. The survey was conducted in three medical universities of China in 2014. A total of 183 third- and 221 fourth-year medical students were included. Microsoft Excel 2007 and SPSS 17.0 were used to perform data extraction and analysis. Results: The survey yielded 500 questionnaires with an 81% response rate. The average blank item in a total of 31 items was 0.96% of which the item that was most often blank was ‘What will happen when a medical error occurs?’. The study population consisted of 143 male and 262 female students. Their attitude to good patient safety was positive, but the students had little knowledge of patient safety especially regarding medical errors and how to report them,. There were no statistical differences across the three universities or grades. The only statistical difference was for the item ‘I would like to discuss with others when I made medical error’ between genders. Conclusions: The results suggest that the knowledge and skill of patient safety from medical students is very poor in China. The patient safety culture should be established properly. Cochrane Database Syst Rev Suppl 1–327 (2015) 214 P1.123 P1.126 Uptake of systematic reviews and meta-analyses in health policy briefs: an analysis of references Development of clinical research support system for conditional coverage with evidence development Wang Q1 , Chen TY2 , Ding HF2 , Zhang H3 , Sun HH2 , Zhang JY4 , Li N2 , Yao L1 , Wang XQ1 , Chen YL1 , Yang KH1 1 Evidence-based Medicine Center of Lanzhou University;Key LaboratoryofEvidence-basedMedicineandKnowledgeTranslation of Gansu Province, China; 2 The Second Clinical Medical College of Lanzhou University, China; 3 School of Basic Medical Sciences of Lanzhou University, China; 4 School of Public Health of Lanzhou University, China Son SK1 , Park DA1 , Lee MJ1 , Lee NR1 , Lee J1 , Yoon JE1 , Lim SW1 1 National Evidence-Based Healthcare Collaborating Agency, South Korea Background: The achievement of universal and equitable access to healthcare, one of the health-related Millennium Development Goals (MDGs), is most likely to be realized through evidence-based health policies and actions. Evidence-based health policy briefs are widely used by WHO and health sectors of many countries, as the most important way of presenting evidence to health policy-makers. They provide potential options based on the best global evidence (such as systematic reviews) and local evidence on priority issues. Currently, there is no analysis of uptake of systematic reviews and meta-analyses in the health policy briefs. Objectives: To explore the uptake of systematic reviews and meta-analyses in the health policy briefs based on the analysis of references. Methods: Two researchers independently handsearched the Health System Evidence (HSE) and World Health Organization (WHO) databases before December 2014 and included health policy briefs in English and Chinese. Two reviewers independently completed data extraction and resolved disagreement by discussion. Results: We included 106 health policy briefs in WHO and HSE from 2001 to 2014. All the policy briefs were published in English; 96% of policy briefs (102) reported the references and the total number was 5683 (range: 6 to 247, mean: 56). The main three sources were: websites (2315, 41%), journals (2165, 38%), and books (60, 1%). The main four types of references were: primary studies (3379, 59%), systematic reviews (317, 6%), clinical practice guidelines (29, 0.5%) and overviews of systematic reviews (14, 0.2%). Among these cited systematic reviews, 172 were Cochrane Reviews and 145 were non-Cochrane reviews. We did not find any network meta-analysis. Conclusions: Most health policy briefs reported references, but the proportion of uptake of systematic reviews was low. We think health brief developers should propose the evidence-based policy options on the basis of up-to-date and high-quality systematic reviews and meta-analysis. As a next step, we will explore the details of cited systematic reviews on the topic and quality and how many policy options were supported the systematic reviews. Background: Since 2014, the Korean government has introduced Selective Reimbursement System to intensify national health coverage for severe diseases including cancer, cardiac, cerebrovascular, and rare diseases. Unfortunately, in most cases, available scientific evidence that can be used as an evidence-base for making policy decision is often limited. Objectives: The purpose of this study was to suggest a model for a clinical research support system for conditional coverage with evidence development. Methods: A search for relevant web pages, a quick literature review, and an in-depth interview with experts were performed for this study. Results: Criteria for prioritization of target technologies, administration of research process, and funding source should be developed to establish evidence development for conditional coverage through clinical research. With respect to the prioritization, the technologies that need additional evidence due to uncertainty of safety and effectiveness and have a high potential of being able to obtain evidence within a short period should be proritized. In the administration of research process, an independent and professional agency including a ‘steering committee’ or ‘sub-committee’ that can provide scientific advice and methodological guidance should manage the clinical research processes. Funding issues are key factors to be considered carefully in order to settle the clinical research support system for conditional coverage with evidence development in smoothly. Principally, it is reasonable that the government should fund the study, as it is in accordance with government policy support. Additionally, various funding sources (e.g. research foundation etc) are necessary in terms of system sustainability. Conclusions: To guarantee a high quality of evidence obtained through the clinical research, the importance of process management must be acknowledged. The actual policy should be designed through the consultation of stakeholders and agencies related to the system, roles that should be performed by authorities and the targeted technology. P1.127 Is evidence-based medicine incorporated into the analysis of European health systems? Decsi T1 , Endrei D2 , Mihályi K3 , Lohner S1 , Boncz I4 1 Hungarian Branch of the German Cochrane Centre, University of Pécs, Pécs, Hungary; 2 Clinical Centre, University of Pécs, Pécs, Hungary; 3 Department of Paediatrics, University of Pécs, Pécs, Hungary; 4 Faculty of Health Sciences, University of Pécs, Pécs, Hungary Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Cochrane Database Syst Rev Suppl 1–327 (2015) 215 Background: There is an increasing requirement for policy makers to ensure that their decisions are evidence-based. In turn, researchers should conduct research addressing key aspects of health policy questions, in order to render possible an effective information transfer between researchers and decision makers as well as an evidence-based approach to policy making. Objectives: The aim of the present study was to investigate the extent to which the idea of evidence-based medicine (EBM) is incorporated into the documents describing the functioning of European healthcare systems. Methods: An electronic search was conducted in the documents available in the European Observatory on Health Systems and Policies database, using the text word ‘evidence’ and the total number of the text word ‘evidence’ referring to EBM was counted. These representations were then classified into one of the following categories: (1) resources of health care; (2) health technology assessment; (3) organisation of health care; and (4) environment of health care within the society. Results: Altogether thel documents mentioned the idea of EBM 413 times; however, there were large differences in both the absolute number of mentions and the incidence of mentioning per page of the document. Two-hundred and eight representations of the idea were assigned to one of the four categories listed above. EBM was mentioned 57 times in connection with resources of health care, 31 times with health technology assessment, 56 times with organisation of health care and 64 times in relation to the social environment of health care. Conclusions: The results of the present study indicate that the methodology of EBM has already been widely used within the documents describing health systems and policies in Europe. However, considerable differences in the extent and depth of applying EBM methodology are detectable in descriptive European health policy documents. to the GRADE principles, taking into account the balance between desirable and undesirable effects, quality of evidence, values and preferences, and costs. Objectives: To assess the possibility of adding recommendations to the Cochrane-derived EVDs and to compare the formulation of the recommendations with the guideline contents. Methods: Treatment of menorrhagia, a common condition in primary care, was chosen as the indicator guideline. We assessed the treatment chapter of the guideline to find whether it was possible to add a recommendation to every Cochrane-based EVD linked to the chapter. The effect of the recommendation on the usability of the guideline was evaluated. Results: Altogether 13 Cochrane-based EVDs were linked to the guideline. The Cochrane Reviews were published in 2000 to 2014 and all the EVDs referred to the latest published version. One EVD already included a recommendation. According to the evaluators, a further eight EVDs could be complemented with a recommendation. In the remaining four EVDs the addition of a recommendation was judged not to be feasible. The results will be presented in table format explaining the basis of these judgments. Conclusions: Recommendations derived from concise evidence summaries and linked to the most essential or critical parts of a point-of-care guideline can provide valuable basis for clinical decision-making. P1.129 Access of evidence-based online databases among main and allied health professionals: a nationwide survey in Taiwan Chiu Y1 , Kuo K1 , Weng Y2 , Chen C1 , Chen K1 , Lo H1 1 Taipei Medical University, Taiwan; 2 Chang Gung Memorial Hospital, Taiwan P1.128 From evidence to recommendation: Cochrane-based evidence summaries as sources of recommendations in a point-of-care guideline Alenius H1 , Jousimaa J1 , Teikari M1 , Kunnamo I1 1 Duodecim Medical Publications Ltd, Finland Background: Evidence-Based Medicine Guidelines (EBMG) is a collection of concise point-of-care guidelines, originating in Finland and published in a number of different language versions. The guidelines are backed up by over 4000 Evidence Summaries (EVDs), of which over 50% are based on Cochrane Reviews. The level of evidence (LoE) in the EVDs is graded according to the principles of the GRADE (Grading of Recommendations Assessment, Development and Evaluation) group to be either high, moderate, low, or very low (indicated by the letters A, B, C and D, respectively). An increasing number of EVDs also include a recommendation formulated by the EBMG editors. The Strength of Recommendation (SRec) is graded to be either strong or weak for or against an intervention according Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Background: There are a variety of evidence-based online databases, but few studies have examined whether main and allied health professionals prefer different resources. Objectives: The current study was to investigate the preference in accessing the online databases between main and allied health professionals of regional hospitals in Taiwan. Methods: A constructed questionnaire survey was conducted from January through April 2011 in nationwide regional hospitals of Taiwan. Questionnaires were mailed to main professionals (physicians and nurses) and allied professionals (pharmacists, physical therapists, technicians, and others), with 6160 valid returns collected. Results: The most commonly-used databases for physicians were MEDLINE/PubMed, the Cochrane Library, UpToDate, and MD Consult. In addition, the commonly-used databases for nurses and technicians were databases in Chinese, including the Index to Chinese Periodical Literature, the National Digital Library of Theses and Dissertations in Taiwan, and the Chinese Electronic Periodical Service. For pharmacists, the most commonly-used database was Micromedex. Overall, the most common motivation for accessing the online databases was self-learning, followed by clinical practice, positional promotion, class assignment, Cochrane Database Syst Rev Suppl 1–327 (2015) 216 instruction preparation, research, medical accreditation and insurance issues. Conclusions: This national survey depicts the pattern of access to the online databases among various health professionals. There were significant differences between and within main and allied health professionals in their use of online databases. The data provide clinical implications for strategies to promote the accessing of evidence-based information. Knowledge translation and communicating evidence P1.130 Filtering the information overload from systematic reviews with ‘Summary of findings’ tables Conway A1 1 HRB-TMRN and NUI Galway, Ireland Background: This presentation will evaluate the ‘Summary of findings’ (SoF) table as a knowledge translation tool to communicate key evidence from systematic reviews to clinicians. Systematic reviews are currently viewed as the highest levels of evidence to support decision-making in healthcare. However, barriers such as time limitations, complexity and large volumes of research can impact negatively on the ability and willingness of potential review users to engage with full versions of systematic reviews. SoF tables may filter the information overload for clinicians enabling them to make better decisions. Objectives: To evaluate the effectiveness of SoF tables on user understanding of findings from systematic reviews. Methods: This systematic review will include published and unpublished, randomised trials and non-randomised trials, where the effects of exposure to SoF tables of systematic reviews on one or more outcomes is measured. The included trials will compare SoF tables to an alternative method of dissemination of systematic reviews. Results: To be confirmed. Conclusions: If the results indicate that SoF tables support increased understanding, then the review will benefit potential users of systematic reviews such as clinicians, guideline developers, policy makers and other stakeholders such as charitable organisations and individuals or groups who inform the patient population and/or the public. It may provide recommendations in a form in which they can quickly access the key findings of future reviews. It may also support these users in making decisions about whether creating SoFs would be a good way of disseminating review findings (and potentially other research findings) within their own organisations. P1.131 Moving from research to practice: training students in translating and communicating evidence Junqueira D1 1 Universidade Federal de Minas Gerais, Cochrane Adverse Effects Methods Group, Brazil Background: Despite global calls for promoting the use of research for informing decisions about health, the use of evidence in policymaking and practice remains a challenge. Objectives: This research aimed (1) to design knowledge translation products to guide informed and shared clinical decisions, and (2) to develop core competences, skills and behaviors in evidence-based concepts for informed decisions about health in a group of pharmacy undergraduate students. Methods: Pharmacy students from Brazil developing their undergraduate thesis were supervised to work on the subjects ‘health education’ and ‘evidence-based practice’ by designing a knowledge translation product to inform the use of evidence, considering local resources and costs. Students had no basic training in research and followed a structured tutorial of activities and learning materials focused on tools for evidence-informed health care. Results: Four undergraduate theses were developed in different topics: (1) metformin for polycystic ovary syndrome, (2) treatment of sinusitis in children, (3) antibiotics for acute otitis media in children, and (4) beta-blockers for hypertension during pregnancy. All of them were based on a Cochrane Systematic Review and included a leaflet or infographic diagram as a final product. During the oral presentation of the theses, students were able to demonstrate critical understanding of the concepts for informing decisions about health in order to influence future professional environments. Conclusions: Undergraduate training of differentiated professional who value the impact that information and health education have for the success of health interventions. Knowledge translation to consumers, healthcare professionals, or policy makers may significantly improve the curriculum of courses in the health area and play an important role in designing expert contents to inform the translation of research into practice and policy. Similar experiences could support Cochrane activities developed to support the use of evidence-based research by health professionals and consumers. P1.133 Treatment guidelines in specialist mental healthcare: how important are they and how to translate them into practice? Fiedler I1 , Girlanda F1 , Becker T1 , Barbui C2 , Kösters M1 1 Ulm University, Ulm, Germany; 2 Department of Public Health and Community Medicine, Section of Psychiatry, University of Verona, Italy Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. Cochrane Database Syst Rev Suppl 1–327 (2015) 217 Background: Clinical practice guidelines have become a standard tool in the attempt to increase the quality of clinical care. They provide a guide to best practice and a framework within which clinical decisions can be made. Nevertheless, knowledge on how to best implement guidelines into mental healthcare is sparse. Objectives: To investigate the efficacy of guideline implementation on provider performance and patient outcomes in mental healthcare settings, and to explore how guidelines should be translated into practice. Methods: We performed a systematic review of randomised controlled trials (RCTs), controlled clinical trials and before-and-after studies comparing guideline implementation strategies versus usual care, and different guideline implementation strategies in patients with severe mental illness recruited in mental healthcare settings. We conducted a meta-analysis of RCTs only. Results: Eighteen studies met our inclusion criteria. Overall, a minority of studies showed a positive statistically significant effect of guideline implementation on provider performance or patient outcomes, and often these studies employed a non-randomised design. The meta-analysis found no statistically significant effect of guideline implementation in terms of provider adherence to guidelines (odds ratio (OR) = 1.25, 95% CI 0.76 to 2.07; P value 0.38), whereas a statistically significant trend in favour of guideline implementation on patient outcomes was revealed (OR = 1.51, 95% CI 1.03 to 2.22; P value 0.04). The results did not show an advantage of guideline implementation strategies performed on professional and organisational level over strategies that only included professional interventions. Conclusions: Current evidence is not sufficient to draw firm conclusions on the beneficial effects of guideline implementation on provider performance and patient outcome. Regarding the limited evidence, we were not able to determine the best way to translate guidelines into practice. In the light of numerous guidelines developed, there is a strong need for further research on this topic by applying optimised methods for the evaluation of guideline implementation. P1.134 Bunches of evidence: displaying ‘Summary of findings’ tables visually as ‘evidence flowers’ Jordan J1 , Babatunde O1 , van der Windt D1 1 Research Institute for Primary Care & Health Sciences, Keele University, United Kingdom Background: ‘Summary of findings’ tables in Cochrane Reviews can be difficult to interpret, particularly by patients and clinicians, and this may limit their use in clinical practice. Evidence flowers were originally developed as a means of expressing the results of large evidence syntheses to stakeholder groups in two research projects. In this presentation we explore the feasibility of summarising the information from ‘Summary of findings’ tables in the form of evidence flowers. Objectives: To suggest a pictorial, simple and novel method of displaying the information in Copyright c2015 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. ‘Summary of findings’ tables of Cochrane Reviews to improve their accessibility to mixed audiences. Methods: Five recent Cochrane Reviews that contain ‘Summary of findings’ tables were purposively selected to represent a range of types of reviews and findings, and evidence flowers were created to represent their ‘Summary of findings’ tables. The petals of the evidence flowers were coloured according to the Grading of Recommendations Assessment, Development and Evaluation (GRADE) evidence rating system and narrative summary of the quantitative information on the petal was kept brief and written in plain language. Effect sizes were indicated with a three-star grading system. Results: The evidence flowers for the five reviews will be presented alongside the corresponding ‘Summary of findings’ tables. As the evidence in ‘Summary of findings’ tables is rated using the GRADE system, it was straightforward to colour the petals of the flower. Deciding on appropriate narrative summary and how to present the quantitative information was more problematic. Agreement on a standardised system for this is needed. Conclusions: Evidence flowers are a novel and visually stimulating method for presenting research evidence. We propose that they are used in addition to ‘Summary of findings’ tables, making the evidence more accessible to a variety of different audiences. P1.135 Advocating for Cochrane evidence: packaging Cochrane HIV/AIDS summaries Oliver J1 , Kredo T1 , Galloway M1 1 South African Cochrane Centre, South African Medical Research Council, South Africa Background: Advocating for evidence is a core approach for Cochrane Strategy 2020. The South African Cochrane Centre (SACC) aims to increase awareness of relevant Cochrane HIV/AIDS Reviews. The Cochrane Library may be difficult for consumers to access due to limited internet availability, and navigating between reviews is time-consuming. Objectives: To promote the use of Cochrane evidence by producing a booklet of plain language summaries for the prevention, testing and treatment of HIV/AIDS, and disseminating it in the African region. Methods: We searched the Cochrane Library to identify all HIV/AIDS reviews published between 2010 and 2014. We used the World Health Organization (WHO) Guidelines for the Prevention and Treatment of HIV as a guide to categorize the interventions. The booklet was produced in print and electronic format (available on the SACC website). The electronic version was disseminated to SACC contacts on and around World AIDS Day 2014 and is ongoing. Stakeholders were informed that printed copies were available. Results: Our search of the Cochrane Library retrieved 58 HIV reviews (14 on prevention; seven on testing and counselling; 30 on treatment including adherence; three on nutritional support; and four on health services and care). As part of our promotion strategy we initially emailed the booklet to 1462 SACC contacts on World AIDS Day 2014 and it was included on the Cochrane Library website. Our stakeholders included Cochrane Database Syst Rev Suppl 1–327 (2015) 218 provincial government HIV/AIDS departments, the National AIDS Council, non-governmental and AIDS tre