Hospital Garrahan
Transcription
Hospital Garrahan
Success Story Hospital Garrahan The Hospital de Pediatría S.A.M.I.C. “Prof. Dr. Juan P. Garrahan” is a highly specialized pediatric institution, leader in Argentina's public health sector. The Hospital registers over 400,000 consultations and 10,000 surgeries per year, and approximately 22,000 children are discharged annually. It is located in the Southern part of the City of Buenos Aires, and includes a Comprehensive Treatment Center for Cancer Patients, eighteen operating rooms, a Transplant Unit, a Neonatology Unit and a Burnt Unit. Flux IT has been working in collaboration with the Cell Biology and Retrovirology Lab. The lab focuses on the study of HIV infection in children. What did the client need? HIV is characterized by a high mutation rate, resulting in great genetic diversity. In order to study the changes in the virus's genome, HIV DNA sequences are obtained from blood samples of HIV-infected people. These sequences must in turn be interpreted by means of specialized analytics software that either translates the DNA sequences into protein sequences or that applies evolution models to predict the impact of the virus genetic changes on the disease evolution, and the patient's response to specific treatments. For doctors to chose the best possible treatments, HIV-infected patients are now required to undergo drug-resistance testing to determine whether their virus has mutated in the presence of antiretroviral drugs. When we were first approached, this task implied interpreting the DNA sequences using on-line tools, and manually loading all observed mutations into the lab's results system. HIV DNA sequences consist of a combination of four letters only (A, C, T, and G), and can have up to 9,000 characters in length. These facts entailed the need to store the virus's genetic data in a particular format (FAS files) and separately from the patient's genetic data. Flux IT with Hospital Garrahan. All rights reserved. 2015. 01/05 Other factors that had to be taken into consideration were the growing number of sequences available, the need to provide easy interpretation of the data, and the required ability to link the viral data to the patient's clinical and lab history data, which could have an influence on the evolution of the disease. The bioinformatics platform SISGEN was created in order to respond to these needs. How did we do it? Methodology When we first started working on SISGEN, the future platform users at the Hospital hadn't yet defined the exact scope of the tool, and our development team had no experience in the field at hand. We were filled with many uncertainties, but thanks to our agile methodologies, we were able to work passed them and succeed. We began working on a first stage of “Analysis, Conceptualization, and Initial Prototyping,” which allowed us to close the existing gap of speaking different languages. Through the initial prototyping of certain features we were able to gain an understanding of the HIV biology and genetics, and also to begin defining the scope of the project. This stage included gathering information, understanding the users profiles [Fig. 1] (thanks to the collaboration and help of usability and user experience analysts), making low-fidelity mockups, drawing entity-relationship diagrams, etc. Fig.1. Analysis of several scenarios based on different users profiles As we collected more information, we developed a Visual Story Map [3], which allowed us to visualize and define the scope of the application [Fig. 3]. Flux IT with Hospital Garrahan. All rights reserved. 2015. 02/05 Sequences Analysis SNPs Analysis Treatments' Data Patients' Data Lab Tests Results Analysis Users Management Load sequences Tag sequences Manage genes and polymorphisms Load treatments Load treatments Load/search lab tests results Manage users, roles and permissions Link to mutation data (jsierra) Filter by sequences Load a patient's SNPs Filter by treatments Filter by patients' data Filter by lab tests results Security and login Locate parts of the virus (locator) Export sequences in FASTA format Filter by SNPs Visualize evolution in time Generate resistance report Fig.2. High-level Story Map We were finally able to clearly define the purpose of the project: “To build a tool that would allow users to easily link clinical data to sequential, treatment and SNPs data; which today are all stored in different media. To build a tool that would simplify the data loading process, that would allow doing searches based on different criteria, and that would provide a user-friendly visualization of the results.” This analysis and conceptualization stage, called Sprint Zero in Agile Methodologies, was a crucial preparation step for the development work. It allowed us to understand the complexity of the project, to develop confidence, and to build a sense of teamwork; much needed attributes to move onto the development stage. We then started working with Scrum. We first built a Product Backlog using our initial Story Map, and followed the steps suggested by this framework. During the Sprint Review at the hospital, where all parties involved were present, users were able to interact with the application and try the newly added features. It was their turn to accept these features, suggest changes or modifications, spot weaknesses, and question whatever was not clear. Their feedback was then used to plan our next sprint. “I think we both learnt a lot from each other. The work methodology applied by the development team at Flux IT and their way of recording the work progress were very agile, well-organized, and easy to follow. At each meeting we identified what we felt as weak points, and continued working on the strong features of the tool”. Paula Aulicino, Biochemist, PhD; University of Buenos Aires. Flux IT with Hospital Garrahan. All rights reserved. 2015. 03/05 What did we do? The Solution The resulting design and architecture (five apparently unrelated modules: Sequences, SNPs, Treatments, Laboratory, and Patients) [Fig. 3] made it possible for each user at the lab to do their own research and follow-ups according to their specific approaches. By means of combined searches, SISGEN allows users to combine any type of data to create new lines of research and begin to address issues unthought-of before, such as being able to retrieve genetic data based on epidemiological, treatment and/or clinical evolution criteria. It also offers the possibility of making comprehensive patient follow-ups by combining different approaches. Fig.3. Data Architecture. SISGEN-HIV does not only provide genetic data storage in a user-friendly format, but it is actually a unified database where HIV genetic sequences, sample-related data, supplementary information from public databases, patient-specific genetic data (Single Nucleotide Polymorphisms, quantitative studies, haplotypes, and HLA gene studies), personal data (from the Hospital's system), and medical history data (diagnoses, a history of perinatology data, treatments, test results) are all stored together. The application provides easy access to all of this information, retrieving useful data and generating relevant results quickly and efficiently. Flux IT with Hospital Garrahan. All rights reserved. 2015. 04/05 Benefits SISGEN-HIV is a highly innovative, added-value product, that helps researchers improve and optimize their work. It provides a secure and efficient way to store important data, and to retrieve valuable information for the formulation of research hypotheses and the production of new knowledge. The development of this tool is yet another proof that the Hospital Garrahan is a national benchmark, having a positive impact on the society. SISGEN also simplifies the creation of drug-resistance testing reports. By simply loading a nucleotide sequence, the application automatically retrieves the mutation and inhibitor data, and creates the drug-resistance report, which can then be directly sent to the Hospital's system or downloaded in document format with a simple click of the mouse. This automatic process basically reduces operation time by half, and minimizes the amount of errors related to manual data loading. SISGEN also guarantees confidentiality regarding patients' information, since the system is installed in the Hospital's own server. The innovative aspect of this tool resides in the interconnection between clinical, genetic and molecular data. When studying different diseases, all these types of data are available and can be analyzed as a whole. This characteristic will help doctors in their decision-making process, and when providing guidance and support to patients and family members. It will also be useful in the search of the immunogenetic or pharmacogenetic factors involved in patients' differences regarding their clinical diagnosis and treatment of diseases. The first version of SISGEN is completely functional for loading patients' sequences and data. It works in communication with the Hospital's database, and it has reduced processing times to half of what they used to be. SISGEN is used on a daily basis. Old Excel and Access databases have been replaced, and the sequencing results have been directly loaded onto SISGEN. “Even when there is still one module missing, which will include lab data, this first version of SISGEN allows us to load patients' sequences and data. It works in communication with the Hospital's database, which simplifies the search of patients and the creation of drug-resistance reports, thus avoiding errors related to manual data loading and reducing processing times to half of what they used to be. We use SISGEN on a daily basis; have replaced old Excel and Access databases, and loaded all sequencing results directly to SISGEN.” Paula Aulicino, Biochemist, PhD; University of Buenos Aires. Flux IT with Hospital Garrahan. All rights reserved. 2015. 05/05