Print this article - Canadian Journal of Public Health

Transcription

CANADIAN JOURNAL of PUBLIC HEALTH
REVUE CANADIENNE de SANTÉ PUBLIQUE
November/December 2012 Volume 103(Supplement 3)
novembre/décembre 2012
Canadian Evidence on
Built Environment and Health
ESTABLISHED IN 1910
cjph.cpha.ca
ÉTABLI EN 1910
Acknowledgements
Production of this special Supplement has been made possible through financial or in-kind
contributions from Health Canada, through the Canadian Partnership Against Cancer’s
CLASP initiative, the Canadian Institute for Health Information, the Canadian Institute
of Planners, and the Heart and Stroke Foundation.
We also would like to thank the organizing partners of the workshop titled, Sharing Knowledge –
Building Links – Advancing Research, Policy and Practice on the Built Environment, held in Ottawa, March
7-9, 2011. Event partners include: the Heart and Stroke Foundation, the Canadian Partnership Against
Cancer, the Canadian Institute for Health Information, the Canadian Institutes of Health Research,
and the Public Health Agency of Canada.
The March 2011 Sharing Knowledge workshop is where the idea of a journal supplement on Canadian
built environment and health research first emerged. Further support by the Canadian Institute for
Health Information, the Canadian Institute of Planners, the Heart and Stroke Foundation, and the
Healthy Canada by Design CLASP initiative through Canadian Partnership Against Cancer’s funding
helped to turn this idea into action.
For access to various resources complementary to this Supplement, such as a series of fact sheets
summarizing the latest Canadian built environment and health research in lay language, visit the
Canadian Institute of Planners Healthy Communities website:
www.cip-icu.ca/health
The views expressed in this Supplement and related materials represent the views of the authors and
do not necessarily represent the views of the project funders.
Canadian Evidence on
Built Environment and Health
TABLE OF CONTENTS
S3
Built Environment Health Research: The Time Is Now for a
Canadian Network of Excellence
N. Muhajarine
S5
Coming to Consensus on Policy to Create Supportive Built
Environments and Community Design
K.D. Raine, N. Muhajarine, J.C. Spence, N.E. Neary, C.I.J. Nykiforuk
S9
Physical Activity Patterns of Children in Toronto: The Relative Role
of Neighbourhood Type and Socio-economic Status
M.R. Stone, G.E. Faulkner, R. Mitra, R.N. Buliung
S15 Linking Childhood Obesity to the Built Environment: A Multi-level
Analysis of Home and School Neighbourhood Factors Associated
With Body Mass Index
J.A. Gilliland, C.Y. Rangel, M.A. Healy, P. Tucker, J.E. Loebach,
P.M. Hess, M. He, J.D. Irwin, P. Wilk
S22 Smart Cities, Healthy Kids: The Association Between Neighbourhood
Design and Children’s Physical Activity and Time Spent Sedentary
D.W. Esliger, L.B. Sherar, N. Muhajarine
S29 Walkable for Whom? Examining the Role of the Built Environment
on the Neighbourhood-based Physical Activity of Children
K. Loptson, N. Muhajarine, T. Ridalls and the Smart Cities, Healthy
Kids Research Team
S35 There’s No Such Thing as Bad Weather, Just the Wrong Clothing:
Climate, Weather and Active School Transportation in Toronto,
Canada
R. Mitra, G. Faulkner
S42 Safe Cycling: How Do Risk Perceptions Compare With Observed
Risk?
M. Winters, S. Babul, H.J.E.H. Becker, J.R. Brubacher, M. Chipman,
P. Cripton, M.D. Cusimano, S.M. Friedman, M.A. Harris, G. Hunte,
M. Monro, C.C.O. Reynolds, H. Shen, K. Teschke
S48 Associations Between Children’s Diets and Features of Their
Residential and School Neighbourhood Food Environments
A. Van Hulst, T.A. Barnett, L. Gauvin, M. Daniel, Y. Kestens, M. Bird,
K. Gray-Donald, M. Lambert
S55 Physical Activity and Nutrition Among Youth in Rural, Suburban
and Urban Neighbourhood Types
C. Shearer, C. Blanchard, S. Kirk, R. Lyons, T. Dummer, R. Pitter,
D. Rainham, L. Rehman, C. Shields, M. Sim
S61 Creating Neighbourhood Groupings Based on Built Environment
Features to Facilitate Health Promotion Activities
D. Schopflocher, E. VanSpronsen, J.C. Spence, H. Vallianatos,
K.D. Raine, R.C. Plotnikoff, C.I.J. Nykiforuk
S67 Examining Aspects of the Built Environment: An Evaluation of a
Community Walking Map Project
C.I.J. Nykiforuk, L.M. Nieuwendyk, S. Mitha, I. Hosler
© Canadian Public Health Association, 2012. All rights reserved.
CANADIAN JOURNAL OF PUBLIC HEALTH • NOVEMBER/DECEMBER 2012
S1
EDITORIAL OFFICES • BUREAUX DE LA RÉDACTION
SCIENTIFIC EDITOR/RÉDACTEUR SCIENTIFIQUE
Gilles Paradis (McGill University; Institut national de santé publique du Québec)
ASSOCIATE EDITORS/RÉDACTEURS SCIENTIFIQUES ADJOINTS
Environmental Health/Santé environnementale
Pierre Ayotte (Université Laval)
Qualitative Research/Recherche qualitative
Joan Eakin (University of Toronto)
Book Reviews/Recensions
Jacqueline C. Gahagan (Dalhousie University)
Intervention Research/Recherche d’intervention
Penny Hawe (University of Calgary)
History of Public Health/Histoire de la santé publique
Maureen Malowany (McGill University)
Quantitative Research/Recherche quantitative
Gilles Paradis (McGill University)
Public Health Practice/Pratique en santé publique
Stéphane Perron (Agence de la santé et des services sociaux de Montréal)
Aboriginal Health/Santé autochtone
Jeffrey L. Reading (University of Victoria)
Communicable Diseases/Maladies transmissibles
Robert S. Remis (University of Toronto)
Public Health Ethics/Éthique en santé publique
Jason Scott Robert (Arizona State University)
Supplements/Suppléments
Reg Warren (University of Toronto)
Global Health/Santé mondiale
Christina Zarowsky (University of the Western Cape)
EDITORIAL BOARD/COMITÉ DE RÉDACTION
Jane Buxton (University of British Columbia)
Tim Caulfield (University of Alberta)
Joanna Cohen (Ontario Tobacco Research Unit)
Paul A. Demers (Cancer Care Ontario)
Debbie Feldman (Université de Montréal)
Lise Gauvin (Université de Montréal)
Diana Gustafson (Memorial University of Newfoundland)
Duncan Hunter (Queen's University)
Igor Karp (Centre de recherche du Centre hospitalier de l’Université de
Montréal)
Nancy Kreiger (University of Toronto)
Benoît Lévesque (Institut national de santé publique du Québec)
Amee Manges (McGill University)
Doug Manuel (Ottawa Hospital Research Institute)
Yang Mao (Public Health Agency of Canada)
Peggy McDonough (University of Toronto)
Nazeem Muhajarine (University of Saskatchewan)
Cameron Mustard (Institute for Work and Health)
Eric Mykhalovskiy (York University)
Jennifer O'Loughlin (Centre de recherche du Centre hospitalier de
l’Université de Montréal)
David Patrick (British Columbia Centre for Disease Control)
Louise Potvin (Université de Montréal)
Amélie Quesnel-Vallée (McGill University)
Bruce Reeder (University of Saskatchewan)
Duncan Saunders (University of Alberta)
Jeannie Shoveller (University of British Columbia)
Colin Soskolne (University of Alberta)
Kay Teschke (University of British Columbia)
BUSINESS OFFICE/SIÈGE SOCIAL
Canadian Journal of Public Health / Revue canadienne de santé publique
404-1525 avenue Carling Avenue, Ottawa, Ontario, Canada K1Z 8R9
Tel./Tél. : 613-725-3769
Fax/Téléc. : 613-725-9826
E-mail/Courriel : [email protected]
www.cjph.cpha.ca
STAFF/PERSONNEL
Karen Craven (Assistant Editor/Rédactrice adjointe)
Debbie Buchanan (Editorial Assistant/Adjointe à la rédaction)
CIRCULATION & SUBSCRIPTIONS/DIFFUSION et ABONNEMENTS
Tel./Tél. : 613-725-3769, ext./poste 126
E-mail/Courriel : [email protected]
www.cjph.cpha.ca
TRANSLATION/TRADUCTION
Marie Cousineau
COVER/COUVERTURE
Photo credit: John Jutras, Smart Cities, Healthy Kids research team.
Humans have always made their home beside large bodies of water, and
so it is on the banks of the South Saskatchewan River. Saskatoon’s “River
Landing”, built recently but likely conceived a long time ago, provides a
perfect public space for children and adults alike.
The Canadian Journal of Public Health is published every two months by the
Canadian Public Health Association. A subscription to the CJPH is included in
the Association’s membership fee.
La Revue canadienne de santé publique est publiée tous les deux mois par
l’Association canadienne de santé publique. La cotisation de membre de
l’Association donne droit à l’abonnement à la RCSP.
All articles published in this journal, including editorials, represent the
opinions of the authors and do not necessarily reflect the official policy of the
Canadian Public Health Association or the institution with which the author is
affiliated, unless this is clearly specified.
Tous les articles publiés dans cette revue, y compris les éditoriaux, représentent
les opinions des auteurs et ne reflètent pas nécessairement la politique officielle
de l’Association canadienne de santé publique ou de l’établissement auquel
l’auteur est affilié, sauf indication contraire.
CJPH is available in microform from University Microfilms International, Ann
Arbor, Michigan and is abstracted by ProQuest and EBSCO.
La RCSP est disponible sur microfilm auprès de University Microfilms
International, Ann Arbor, Michigan et est abrégée par ProQuest et EBSCO.
Indexed in the Canadian Periodical Index, Index Medicus, Social Science
Citation Index and Repère.
Répertorié dans Canadian Periodical Index, Index Medicus, Social Science
Citation Index et Repère.
Reprints of articles, minimum 50, are available from the business office of the
Journal (price on request). Contents may be reproduced only with the prior
permission of the Editorial Board. A maximum of 30 photocopies of articles are
permitted with acknowledgement of CJPH.
On peut se procurer des tirés-à-part d’article, minimum de 50 exemplaires,
auprès des bureaux de la Revue (prix disponible sur demande). On ne peut
reproduire le contenu de la Revue qu’avec la permission préalable de la
rédaction. Il est permis de copier un maximum de 30 articles à condition de
bien indiquer la source.
PUBLISHER/ÉDITEUR
Canadian Public Health Association/
Association canadienne de santé publique
Debra Lynkowski, LlB, Chief Executive Officer/Chef de la direction
© Association canadienne de santé publique, 2012. Tous droits réservés.
ISSN 0008-4263
S2
REVUE CANADIENNE DE SANTÉ PUBLIQUE • VOL. 103, SUPPLÉMENT 3
Publications Mail Agreement #40062779. Return undeliverable Canadian
addresses to: Circulation Department, Canadian Public Health
Association, 300-1565 Carling Avenue, Ottawa, Ontario, K1Z 8R1, E-mail:
[email protected].
Numéro de convention de la Postes-publications 40062779.
Retourner toute correspondance ne pouvant être livrée au Canada
au : Service des publications, Association canadienne de santé
publique, 1565 avenue Carling, bureau 300, Ottawa (Ontario) K1Z 8R1,
courriel : [email protected].
FOREWORD
Built Environment Health Research: The Time Is Now for a
Canadian Network of Excellence
Nazeem Muhajarine, PhD
E
very generation since World War II has faced its own great public health challenges. In the post-war period, the challenge was
the integration back into society of young men and women
who had fought a war that took a terrible toll on life and spirit; in
the 1960-70s, efforts to control or eradicate smallpox, polio and
malaria dominated; in the 1980-90s, tobacco control, heart disease,
stroke and HIV/AIDS commanded attention; and in contemporary
times, SARS, the H1N1 flu epidemic, mental health and a host of
natural and human-made catastrophes have been added to the mix.
But of all the great public health challenges, the global epidemic of
obesity has emerged as the nemesis of our generation. The numbers
paint a grim picture. Obesity is a disorder in its own right, but more
pervasively it is also the gateway to many other chronic conditions.
The public health and medical care costs attributable to obesity are
staggeringly large, and the personal and social costs are not far
behind. It is against this backdrop that we offer this special supplement on the built environment and health from a Canadian
perspective.
Why this supplement, and why now? While it is clear there is an
urgent need to take action to address obesity, particularly in children, it is also very apparent that the empirical body of evidence
regarding the determinants of obesity – especially those most
upstream, such as the built environment, time use and technology
change – is only now taking shape. The sense of urgency to halt
the childhood obesity epidemic has helped catalyze political
processes in some local jurisdictions aimed at making environmental changes by altering public policy. We need new research,
especially as it relates to Canadian cities, to indicate which policydriven built environmental factors are the most important contributors to obesity, and to understand the mechanisms through
which they work. Such evidence is critically needed to deepen the
policy debate, leading to action with greater promise of decreasing
childhood and adult overweight and obesity in Canada.
The articles in this supplement present current Canadian evidence supporting the impact of the built environment on health,1-11
particularly with regard to child health and obesity. Collectively,
these works represent the contributions of multidisciplinary teams
of researchers from all five regions of Canada and offer evidence
linking various aspects of built and food environments (defined
around neighbourhoods and schools) and community design, and
their impact on active transportation, physical activity, diet and
obesity.
Reports from studies in three Canadian cities – Toronto, Kingston
and Saskatoon – investigate types of urban form (for example, as
one study identified: grid-pattern, mixed grid- and curvilinearpattern, or curvilinear-pattern neighbourhoods) in order to understand their impact on physical activity or BMI.1-3 These papers push
the threshold of current built environment research by going back
to the basics – looking at the design of our urban centres and neighbourhoods, and how that constrains or facilitates people’s choices,
activities and even residential selection. It is necessary to start with
the basic form (structure) of urban and rural neighbourhoods as
that is the blueprint that directs what gets built, as we delve into the
specifics of built characteristics.
Seven of the ten papers in this supplement report on children
between the ages of 8 and 14 years.1-7 Are there any theoretical or
developmental reasons that make children of this age group particularly advantageous to study? Where in neighbourhoods, or
when and how do they accumulate their physical activity? In practical terms, 10-14 year-olds may be an ideal group to study given
that they are old enough to make choices regarding travel and
mobility but not so old as to be completely travel independent
(i.e., driving a vehicle). There may be neurodevelopmental reasons
as well. Between the ages of 11-15, youth undergo a second phase
of brain development specifically related to spatial configuration
and analysis.12 According to environmental psychologists, important cognitive development occurs through the processes of memorizing landmarks and the sequence of routes and through
navigating the integration of routes. Environments that stimulate
this development are ones in which navigation and spatial orientation are challenged and in which opportunities for independent
travel are facilitated. As a society, we should, according to Weston,12
develop cities that allow this to occur in young people.
Many of the studies in this special issue and elsewhere report
physical activity levels, or access to food retailers, in relation to
neighbourhoods, with or without the specification of a surrounding area or buffer zone.1-7 The underlying assumption is that this
activity or access is occurring in and around participants’ home or
school neighbourhoods. These assumptions now need to be tested,
by measuring where, when and what types of activities and food
access occur and under what circumstances (e.g., structured/registered
type of activity; free play). It is likely that the location and types of
activity are distributed differentially across socio-economic status
and, therefore, neighbourhood types as well. We know that frequently children get driven to a structured activity, and when this
Author’s Affiliation
Community Health and Epidemiology, College of Medicine, University of Saskatchewan,
107 Wiggins Road, Saskatoon, SK S7N 5E5, E-mail: [email protected]
S3
FOREWORD
occurs the locations are often outside of their residential or school
neighbourhoods. The apparent inconsistencies seen in some of the
research – for example, greater physical activity in children from
high SES neighbourhoods as well as inner-city neighbourhoods –
could be explained by a careful delineation not only of the intensity of physical activity level but also of where these activities occur
and of what type. We must resist the generalization that what is
good for adults in terms of built environment and health is also
necessarily good for young children. Much of the behaviour
observed in young children is strongly influenced by their adult
caregivers (and their peers); this simple fact has not as yet been adequately accounted for with regard to much of the built environment and children’s physical activity and diet research.
Several of the studies in this supplement have defined built
and/or food environments in relation to the neighbourhoods
where participating children reside and the schools they attend.1-7
This differentiation of environments makes sense given the amount
of time children spend at school compared to at home during weekdays. However, an important finding reported in this issue is the
difference in physical activity levels, for both boys and girls, during
school days compared to on weekends.1,3 The level of physical activity and active living in general for children is not only spatially but
also temporally patterned – within a day, as well as across the week.
The weekday–weekend physical activity levels are different enough,
and consistently so, that future studies are well advised not to treat
all days of the week as equal when physical activity measures are
taken or analyzed. It is increasingly clear that we need to understand what the contextual (including the built environment) and
individual determinants of physical activity for children are on the
weekends, as they may be distinct from the determinants shaping
activity on the school days. It follows then that we need to be more
precise when we consider defining built environments in relation
to schools and residences. Obviously school-based definitions of
built environments may not be relevant when considering weekend
physical activity levels. On the other hand, when considering active
transportation to school, not only is the distance between home
and school important, but further gains in insight are likely to be
made if we are able to link the residential neighbourhood and
school neighbourhood built environments in a seamless manner. In
other words, what children are likely to see outside their homes
and surrounding their schools is as important in influencing their
and their parents’ decisions as what they encounter throughout
the travel path from home to school and back again.
S4
Like the youth who form the subject of many of the papers here,
built environment health research is still a young field, at times awkward, but with much energy and potential. As the supplement
demonstrates, Canadian researchers are making important contributions to this quickly evolving field. There is however much work
yet to do. The next key steps involve creating more clarity in definitions and operationalization of concepts, measurement and integration of multiple methods, and deeper engagement and commitment
for creating a community of researchers in this field. The time is now
for a coordinated national effort in built environment and health
research – a network of centres on built environment. It is through
an escalation of current efforts, integration of local research into a
national network, and engaging of partners across sectors, locally
and nationally, that we will curb childhood obesity in Canada.
REFERENCES
1.
Stone MR, Faulkner GE, Mitra R, Buliung RN. Physical activity patterns of
children in Toronto: The relative role of neighbourhood types and socioeconomic status. Can J Public Health 2012;103(Suppl. 3):S9-S14.
2. Gilliland JA, Rangel CY, Healy MA, Tucker P, Loebach JE, Hess PM, et al. Linking childhood obesity to the built environment: A multi-level analysis of
home and school neighbourhood factors associated with body mass index.
Can J Public Health 2012;103(Suppl. 3):S15-S21.
3. Esliger D, Sherar L, Muhajarine N. Smart Cities, Healthy Kids: The association between neighbourhood design and children’s physical activity and time
spent sedentary. Can J Public Health 2012;103(Suppl. 3):S22-S28.
4. Loptson K, Muhajarine N, Ridalls T and the Smart Cities, Healthy Kids
Research Team. Walkable for whom? Examining the role of the built environment on the neighbourhood-based physical activity of children. Can J
Public Health 2012;103(Suppl. 3):S29-S34.
5. Mitra R, Faulkner G. There’s no such thing as bad weather, just the wrong
clothing: Climate, weather and active school transportation in Toronto,
Canada. Can J Public Health 2012;103(Suppl. 3):S35-S41.
6. Van Hulst A, Barnett TA, Gauvin L, Daniel M, Kestens Y, Bird M, et al. Associations between children’s diets and features of their residential and school neighbourhood food environments. Can J Public Health 2012;103(Suppl. 3):S48-S54.
7. Shearer C, Blanchard C, Kirk S, Lyons R, Dummer T, Pitter R, et al. Physical
activity and nutrition among youth in rural, suburban and urban neighbourhood types. Can J Public Health 2012;103(Suppl. 3):S55-S60.
8. Winters M, Babul S, Becker HJEH, Brubacher JR, Chipman M, Cripton P, et al.
Safe cycling: How do risk perceptions compare with observed risk? Can J
9. Schopflocher D, VanSpronsen E, Spence JC, Vallianatos H, Raine KD,
Plotnikoff RC, Nykiforuk CIJ. Creating neighbourhood groupings based on
built environment features to facilitate health promotion activities. Can J
10. Nykiforuk C, Nieuwendyk L, Mitha S, Hosler I. Examining aspects of the built
environment: An evaluation of a community walking map project. Can J
11. Raine KD, Muhajarine N, Spence JC, Neary NE, Nykiforuk CIJ. Coming to
consensus on policy to create supportive built environments and community design. Can J Public Health 2012;103(Suppl. 3):S5-S8.
12. Weston L. Building cities for young people: Why we should design cities with
preteens and young teens in mind. J Urban Design 2010;15(3):325-34.
© Association canadienne de santé publique, 2012. Tous droits réservés.
COMMENTARY
Coming to Consensus on Policy to Create Supportive
Built Environments and Community Design
Kim D. Raine, PhD, RD,1 Nazeem Muhajarine, PhD,2 John C. Spence, PhD,3 Neil E. Neary, MPH-HP,1
Candace I.J. Nykiforuk, PhD1
ABSTRACT
In April 2011, a conference with invited experts from research, policy and practice was held to build consensus around policy levers to address
environmental determinants of obesity. The gap between existing policy tools and what can promote health through community design is a major
policy opportunity. This commentary represents a consensus of next actions towards creating built environments that support healthy active living. The
policy environment and Canadian evidence are reviewed. Issues and challenges to policy change are discussed. Recommendations to create supportive
built environments that encourage healthy active living in communities include the following: 1) empower planning authorities to change bylaws that
impede healthy active living, protect and increase access to green space, introduce zoning to increase high density, mixed land use, and influence the
location and distribution of food stores; 2) establish stable funding for infrastructure promoting active transportation and opportunities for recreation;
3) evaluate the effectiveness of programs to improve the built environment so that successful interventions can be identified and disseminated;
4) mandate health impact assessment of planning, development and transportation policies to ensure that legislative changes promote health and
safety; 5) frame issues to dispel myths and to promote protection from obesity risk factors.
Key words: Child; adolescent; health status; obesity; health policy; environment design
La traduction du résumé se trouve à la fin de l’article.
I
n April 2011, a consensus conference was held with invited
experts from research, policy and practice fields. The conference
aimed to build consensus around policy levers to address environmental determinants of obesity, including next logical steps
toward further policy action. We identified a range of opportunities
for action, based upon a synthesis of international recommendations, best evidence from the literature and previous prioritysetting within the Canadian policy milieu;1-3 creating supportive
built environments for healthy active living was one opportunity
discussed. We invited conference participants to share with the
group high-level insights from their experience (available political
opportunities/appetite for change) and research (evidence of what
works/does not work). Consensus was reached through facilitated
discussion following presentation of evidence and the environmental context. A consensus paper was drafted and circulated for
revision. This commentary represents a consensus of next actions
towards creating supportive built environments for healthy active
living.
nutrient density foods, limiting access to healthy foods, and exposing people to stressors that may moderate how energy intake and
expenditure relate to weight status. For example, design elements
that include road network patterns leading to increased connectivity and safety, pedestrian accessibility, mixed land use, parks and
playgrounds contribute to the walkability of a neighbourhood and
have a tremendous impact on levels of physical activity, especially
with respect to active transportation and play.5 Residential density
and zoning bylaws contribute to community design and access to
opportunities for both physical activity and healthy eating.
Many Canadian jurisdictions are acknowledging a need for
changes to built environments, including zoning and bylaws, to
promote healthy weights.6 According to the 2011 Active Healthy
Author Affiliations
“The built environment is part of our physical surroundings and
includes the buildings, parks, schools, road systems, and other infrastructure that we encounter in our daily lives”.4 The built environment may influence the behaviour or weight status of populations
by decreasing opportunities for physical activity, increasing opportunities to be sedentary, increasing exposure to high-energy, low-
1. Centre for Health Promotion Studies, School of Public Health, University of Alberta,
Edmonton, AB
2. Community Health and Epidemiology, College of Medicine, University of
Saskatchewan, Saskatoon, SK
3. Faculty of Physical Education and Recreation, University of Alberta, Edmonton, AB
Correspondence: Kim D. Raine, Edmonton Clinic Health Academy #3-291, 1140587th Ave. NW, Edmonton, AB T6G 1C9, Tel: 780-492-9415, Fax: 780-492-0364,
E-mail: [email protected]
Acknowledgements: This research was funded by a Meeting, Planning and
Dissemination grant from the Canadian Institutes of Health Research (CIHR). K. Raine
acknowledges an Applied Public Health Chair Award from the Heart and Stroke
Foundation and CIHR. The authors thank Manuel Arango, Susan Buhler,
Timothy Caulfield, Diane Finegood, Samantha Hartley-Folz, Bill Jeffery, Jane Landon,
Craig Larsen, Tim Lobstein, Lyne Mongeau, Suzie Pellerin, Lisa Petermann,
Monique Potvin Kent, Shandy Reed and Michele Simon for their participation in the
consensus conference.
Conflict of Interest: None to declare.
Policy environment
S5
CONSENSUS ON BUILT ENVIRONMENT POLICY
Kids Canada (AHKC) Report Card,5 “In the past year, there has been
a rise in the number of communities and governments taking this
multi-sectoral/level approach in their strategies to promote child
and youth physical activity and reduce sedentary pursuits by targeting the built environment” (p. 46). Yet, the same 2011 AHKC
report card disclosed that only 16% of municipalities had a formal
transportation master plan and less than 20% of communities
required mandatory safe walking and biking routes in the development of new areas, reconstruction of roads or retrofitting of existing communities. In 2008, the AHKC reported that 96% of
Canadian municipalities had at least one bylaw discouraging physical activity or free play, such as restrictions on skateboarding, street
hockey or bicycling.7 The inherent contradiction is that municipal
bylaws emphasizing safety from injury (safe play) function to
counter promotion of an active lifestyle as part of well-being. The
yawning gap between existing policy tools at the municipal level
and what we acknowledge as increasingly necessary to promote
active lifestyle through community design is a major policy opportunity that needs to be addressed.
Recent surveys of decision-makers from all levels of government,
schools and workplaces in Alberta show support for policy initiatives that address land-use design and zoning to promote active living, but to a lesser degree to promote healthy eating. Respondents
supported implementing transportation policies designed to promote physical activity through safe routes, cycle facilities, adequate
lighting, etc. (89%); enhancing the quantity and quality of green
spaces in all neighbourhoods (93%); changing the design of our
neighbourhoods and communities to encourage informal physical
activity in daily life (85%); and changing building and community design standards to discourage sedentary activities (63%). The
same survey found little support for using land-use decisions and
zoning to promote healthy eating. In fact, a large albeit minority
proportion of respondents (46%) supported zoning to increase the
number of small grocery stores at walking distance from residences
in every neighbourhood, and only 28% supported zoning to limit
fast food restaurants per square kilometre.8
Public opinion, however, is generally much more positive about
such measures, 56% of Albertans surveyed indicating their support
for zoning to limit fast food restaurants per square kilometre, and
74% reporting support for zoning to restrict the supply of junk food
near schools.9 Similarly, 72% of respondents in Quebec claimed
they were supportive, in principle, of restrictions on fast food outlets near schools. The support increases to 83%-87% when the possibility of actually adopting regulatory measures is proposed.10
Evidence
A systematic review has reported evidence for associations between
built environment and diet, physical activity and sedentary behaviour.6 However, while systematic reviews of epidemiologic studies
of the built environment and obesity have found statistically significant associations in approximately half of the studies reviewed,
heterogeneity across studies limits the strength of evidence.11
Looking specifically at Canadian studies, including those in this
supplement, different aspects of the built environment appear to be
related to obesity and its proximal determinants in varied urban
contexts. Specifically, in Toronto, residential density has been associated with decreased population obesity, and in Vancouver, residential density, land-use mix, street connectivity and a composite
S6
walkability index are all associated with decreased population obesity among adults.12 In Ottawa, neighbourhoods with fewer recreation opportunities were associated with higher body mass index
(BMI) among adults.13
The results emerging from built environment studies specifically focusing on children point to a more complex picture. In Saskatoon, children 10-14 years of age living in fractured-grid
neighbourhoods accumulated less physical activity and more
sedentary time per day than those in grid-pattern or curvilinearpattern neighbourhoods.14 Studies from Halifax and Toronto in this
issue further elucidate the potential mitigating effects of neighbourhood level socio-economic status. A study of children 12-14
years of age in Halifax reported that rates of physical activity in
children from schools in lower socio-economic areas were higher in
urban than in suburban or rural settings.15 In Toronto, children
who attended schools in more affluent neighbourhoods had more
positive physical activity profiles across the week.16
The findings on the effect of the built environment on children’s
physical activity and obesity reported in this issue extend the
results from several earlier studies conducted in Canadian cities.
Studies from Edmonton have reported that a walkable neighbourhood design, specifically intersection density, was associated with
decreased childhood obesity among girls.17 Among pre-adolescent
children in Alberta, neighbourhood safety, sidewalks and parks are
negatively correlated with body weight.18 Further, an Ontario study
revealed that children with a park playground within 1 km from
home were five times more likely to be at healthy weight than children without such access.19 Studies have shown that children who
live in neighbourhoods with fewer amenities or lacking neighbourhood access to sidewalks, walking paths, parks or playgrounds,
or recreation or community centres had 20%-45% higher odds of
being obese/overweight.20 These findings are supported by a study
in this issue from London, ON, which reports that public recreation
opportunities within a 500-metre network distance from home
were associated with lower BMI in children.21
Perceptions of built environment are also important. In Edmonton, children aged 6 to 12 described more active transportation in
their neighbourhoods when they lived in highly walkable areas.22
Among adults in Edmonton, neighbourhood choice (for ease of
walking) was associated with a lower BMI, whereas objectively
measured walkability was not a significant influence on BMI.23 Perceptions of traffic as a barrier to walking also predicted higher
BMI.24
Community design, zoning and neighbourhood social factors are
also associated with physical access to foods. For example, Canadian research has found that fast food outlets are more prevalent in
neighbourhoods of lower socio-economic status,25 while supermarkets – sources of wider food choices, including healthy foods –
are less prevalent in these neighbourhoods.26 A Montréal study
reported in this issue found that children attending a school in
neighbourhoods with a higher number of unhealthy relative to
healthy food establishments scored most poorly on dietary outcomes.27 Additionally, a study in Halifax found that dietary quality
was higher among youth in higher than in lower socio-economic
urban settings.15 Further, there was evidence from an Edmonton
study that the shorter the distance to healthier food sources from
one’s residence, the less the likelihood of obesity.28 With regard to
the influence of neighbourhood food environment on healthy eat-
ing among grade 7 and 8 students in London, ON, the proximity
of convenience stores to students’ homes and proximity of schools
to convenience stores and fast food outlets were all significantly
associated with decreased healthy eating index scores.29
Evidence of the impact of the built environment on behaviours
and weight status is growing. There are small but meaningful associations observed between aspects of the built environment and
behavioural determinants/risk of obesity, although the findings are
inconsistent. Perceptions of the environment may be as important
as objectively measured aspects of environments. Over the past four
years, the Heart and Stroke Foundation and Canadian Institutes of
Health Research have made research on obesity and built environment a strategic focus, and several Canadian research projects have
been funded that contribute to the evidence base; several of those
studies are reported in this supplement. Yet, evidence on the effectiveness of built environment interventions is sparse. Very little intervention research has linked the effectiveness of changing
environments to behavioural or weight and health outcomes, especially in a policy context,6 suggesting a serious evidence gap.
Issues and challenges
Changing the built environment significantly is a massive
undertaking with potentially huge financial investment needs
Building transportation infrastructure is a significant investment
that spans jurisdictions’ decision-making structures and involves
sectors not traditionally involved with health. For example, building a road may be primarily a provincial responsibility, yet incorporating bike lanes or increasing public transit may be a municipal
one. Urban planners with municipal mandates may face challenges
as to what they can accomplish, as provincial laws and policies may
be outdated or too restrictive. Physical activity, food access and
health may not yet be a high priority for transportation and planning sectors or even for departments of finance. Indeed, given the
multiple demands on limited municipal tax bases, finding sufficient resources to maintain, change or build infrastructure supportive of health-promoting community design may be a
challenge.
Smaller, localized changes in built environments may provide the
opportunity for incremental changes
Focusing on school environments or municipal bylaws is not only
more manageable politically and practically, but can also provide
preliminary evidence of local effectiveness as exemplars for others.
Public support for change in built environments may need to take
into account parental concerns about child safety
Enhancing safety (i.e., fewer traffic incidents involving pedestrians and cyclists) through improved built environmental design
may have more immediate resonance with parents than changes
in health status or obesity. Social perceptions, such as parental
fears about child safety (i.e., child abduction by strangers, which
is extremely rare), are significant factors that need to be taken
into account when promoting built environment changes. Connecting issues of violence prevention with safety challenges, dispelling myths and fears, and engaging public support for change
should not further exacerbate public fears but, rather, help to dispel them.
Recommendations
To create supportive built environments that encourage physical
activity, active living and access to healthy foods in local
communities, we recommend the following:
• Empower local planning authorities to
– change bylaws that have an effect of restricting physical activity
– initiate programs to help protect and increase access to green
space, including parks and playgrounds
– introduce zoning bylaws that increase high density and mixed
land use
– through a combination of incentives (tax shelters) and constraints (zoning bylaws) influence the location and distribution of food stores, including fast food stores and suppliers of
fruit and vegetables.
• Establish stable, long-term funding for municipal infrastructure
that promotes active transportation and provides opportunities
and facilities for recreation.
• Evaluate the effectiveness of local programs in improving aspects
of the built environment so that successful intervention strategies can be identified and scaled up.
– Evidence of what interventions work under which conditions
is likely to come from natural experiments assumed by committed communities based on the growing evidence associating built environment with obesity and health.
– It is essential for researchers to work with planners and policymakers to capture the impact of local changes.
• Make health impact assessment of planning, development and
transportation policies mandatory to ensure that legislative
changes increase walking, cycling and safety for children and citizens.
– As the impact of the built environment, including land use
and community design, on the health of the population
becomes increasingly evident, promoting the development
and implementation of public policies conducive to health
suggests consideration of the health impacts of new and standing policies in major project planning.
– Development of tools to assist municipalities and provincial
jurisdictions in doing such assessments is required.
• Advocate the framing of issues so as to dispel myths and fears
about child safety (from abduction or injury) and to promote
child safety from obesity risk factors (e.g., physical inactivity,
sedentary behaviours).
CONCLUSION
Consensus around policy levers to address environmental determinants of obesity, including next logical steps toward further policy action, led to concrete recommendations for researchers,
practitioners and policy-makers to create supportive built environments that encourage physical activity, active living and access to
healthy foods in local communities.
REFERENCES
1.
2.
Alberta Policy Coalition for Cancer Prevention. Our Focus [online]. 2011.
Available at: http://abpolicycoalitionforprevention.ca/our-focus.html
(Accessed July 30, 2011).
The Canadian Partnership Against Cancer. Canadian Priorities for Addressing Obesity as a Cancer and Chronic Disease Risk Factor. Toronto, ON: Nutrition and Physical Activity Policy Alignment in Action Initiative, 2010.
S7
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
S8
Heart and Stroke Foundation. Community Design, Physical Activity, Heart
Disease and Stroke. 2012. Available at: http://www.heartandstroke.com/site/
c.ikIQLcMWJtE/b.3820627/k.662E/Position_Statements__Community_
Design_physical_activiy_heart_disease_and_stroke.htm (Accessed June 12,
2012).
Health Canada. Natural and Built Environments. Ottawa, ON: Division of
Childhood and Adolescence, 2002.
Active Healthy Kids Canada. Don’t Let This Be the Most Physical Activity Our
Kids Get After School: The Active Healthy Kids Canada 2011 Report Card on
Physical Activity for Children and Youth. Toronto, ON: Active Healthy Kids
Canada, 2011.
Raine K, Spence JC, Church J, Boulé N, Slater L, Marko J, et al. State of the Evidence Review on Urban Health and Healthy Weights. Ottawa, ON: Canadian
Institute for Health Information, 2008.
Active Healthy Kids Canada. Canada’s Report Card on Physical Activity for
Children and Youth. Toronto, ON: Active Healthy Kids Canada, 2008.
Alberta Policy Coalition for Cancer Prevention. APCCP Survey Results for
Healthy Eating, Active Living, Alcohol Misuse and Tobacco Reduction Issues:
Decision-Maker Survey of Knowledge, Attitudes and Beliefs. Edmonton, AB:
Alberta Policy Coalition for Cancer Prevention, 2009.
Alberta Policy Coalition for Cancer Prevention. Alberta Survey of Knowledge,
Attitudes, and Beliefs. Edmonton, AB: Alberta Policy Coalition for Cancer Prevention, 2010.
Association pour la santé publique du Québec. The School Zone and Nutrition: Courses of Action for the Municipal Sector. Montreal, QC: Association
pour la santé publique du Québec, 2011.
Feng J, Glass TA, Curriero FC, Stewart FC, Stewart WF, Schwartz BS. The built
environment and obesity: A systematic review of the epidemiologic evidence.
Prev Med 2009;16(2):175-90.
Pouliou T, Elliott SJ. Individual and socio-environmental determinants of
overweight and obesity in urban Canada. Health & Place 2010;16(2):389-98.
Prince SA, Tremblay MS, Prud’homme D, Colley R, Sawada M, Kristjansson E.
Neighbourhood differences in objectively measured physical activity, sedentary time and body mass index. Open J Prev Med 2011;1(3):182-89.
Esliger D, Sherar L, Muhajarine N. Smart cities, healthy kids: The association
between neighbourhood design and children’s physical activity and time
spent sedentary. Can J Public Health 2012;103(Suppl. 3):S22-S28.
Shearer C, Blanchard C, Kirk S, Lyons R, Dummer T, Pitter R, et al. Physical
activity and nutrition among youth in rural, suburban and urban neighbourhood types. Can J Public Health 2012;103(Suppl. 3):S55-S60.
Stone MR, Faulkner GE, Mitra R, Buliung RN. Physical activity patterns of
children in Toronto: The relative role of neighbourhood type and socioeconomic status. Can J Public Health 2012;103(Suppl. 3):S9-S14.
Spence JC, Cutumisu N, Edwards J, Evans J. Influence of neighbourhood
design and access to facilities on overweight among preschool children. Int J
Pediatr Obes 2008;3:109-16.
Davidson Z, Simen-Kapeu A, Veugelers P. Neighborhood determinants of selfefficacy, physical activity, and body weights among Canadian children.
Health & Place 2010;16(3):567-72.
Potwarka LR, Kaczynski AT, Flack AL. Places to play: Association of park space
and facilities with healthy weight status among children. J Community Health
2008;33(5):344-50.
Singh GK, Siahpush M, Kogan MD. Neighborhood socioeconomic conditions,
built environments, and childhood obesity. Health Aff (Millwood)
2010;29(3):503-12.
Gilliland JA, Rangel CY, Healy MA, Tucker P, Loebach JE, Hess PM, et al. Linking childhood obesity to the built environment: A multi-level analysis of
home and school neighbourhood factors associated with body mass index.
Holt NL, Cunningham CT, Sehn ZL, Spence JC, Newton AS, Ball GD. Neighborhood physical activity opportunities for inner-city children and youth.
Health & Place 2009;15:1022-28.
23. Berry TR, Spence JC, Blanchard CM, Cutumisu N, Edwards J, Selfridge G. A
longitudinal and cross-sectional examination of the relationship between reasons for choosing a neighbourhood, physical activity and body mass index.
Int J Behav Nutr Phys Act 2010;7(1):57.
24. Berry T, Spence J, Blanchard C, Cutumisu N, Edwards J, Nykiforuk C. Changes
in BMI over 6 years: The role of demographic and neighborhood characteristics. Int J Obes 2010;34(8):1275-83.
25. Hemphill E, Raine K, Spence JC, Smoyer-Tomic KE. Exploring obesogenic food
environments in Edmonton, Canada: The association between socioeconomic
factors and fast-food outlet access. Am J Health Promot 2008;22(6):426-32.
26. Smoyer-Tomic KE, Spence JC, Raine KD, Amrhein C, Cameron N, Yasenovskiy V,
et al. The association between neighborhood socioeconomic status and exposure to supermarkets and fast food outlets. Health & Place 2008;14(4):740-54.
27. Van Hulst A, Barnett TA, Gauvin L, Daniel M, Kestens Y, Bird M, et al. Associations between children’s diets and features of their residential and school
neighbourhood food environments. Can J Public Health 2012;103(Suppl.
3):S48-S54.
28. Spence JC, Cutumisu N, Edwards J, Raine KD, Smoyer-Tomic K. Relation
between local food environments and obesity among adults. BMC Public
Health 2009;9(1):192.
29. He M, Tucker P, Gilliland J, Irwin JD, Larsen K, Hess P. The influence of local
food environments on adolescents’ food purchasing behaviors. Int J Environ
Res Public Health 2012;9:1458-71.
RÉSUMÉ
En avril 2011, une conférence d’experts invités du monde de la
recherche, des politiques et de la pratique a cherché à construire des
consensus autour de leviers politiques pour aborder les déterminants
environnementaux de l’obésité. Le fossé entre les outils stratégiques
existants et ceux qui pourraient favoriser la santé par le design
communautaire présente une importante occasion stratégique à saisir.
Ce commentaire expose le consensus des experts sur les prochaines
étapes en vue de la création de milieux bâtis favorisant une vie active
saine. Nous passons en revue l’environnement politique et les données
probantes canadiennes. Les enjeux et les difficultés des changements
d’orientation sont abordés. Les recommandations en vue de créer des
milieux bâtis qui encouragent une vie active saine dans les communautés
sont les suivantes : 1) habiliter les responsables de la planification à
changer les règlements qui nuisent à une vie active saine, à protéger et
élargir l’accès aux espaces verts, et à introduire un zonage qui accroît la
densité et l’utilisation mixte des sols et qui influence l’emplacement et la
répartition des magasins d’alimentation; 2) établir des budgets de
financement stables pour les infrastructures qui favorisent le transport
actif et les possibilités de loisir; 3) évaluer l’efficacité des programmes
d’amélioration du milieu bâti pour que les interventions fructueuses
puissent être identifiées et disséminées; 4) ordonner l’évaluation des
incidences sur la santé pour toutes les initiatives de planification et les
politiques de développement et de transport, afin que les modifications
législatives favorisent la santé et la sécurité; 5) présenter les enjeux de
manière à déboulonner les mythes et à promouvoir la protection contre
les facteurs de risque de l’obésité.
Mots clés : enfant; adolescent; état sanitaire; obésité; politique sanitaire;
conception de l’environnement
QUANTITATIVE RESEARCH
Physical Activity Patterns of Children in Toronto: The Relative Role
of Neighbourhood Type and Socio-economic Status
Michelle R. Stone, PhD,1 Guy E. Faulkner, PhD,2 Raktim Mitra, PhD,3 Ron N. Buliung, PhD4
ABSTRACT
Objective: A child’s opportunity for physical activity and the safety of engaging in activity are influenced by built environment (BE) elements. This study
examined the relationship of neighbourhood type and socio-economic status (SES) with activity using a sampling frame that purposely located schools
in varying neighbourhoods to ensure that there was variability in BE characteristics and SES.
Methods: Participants (1,027 Grade 5 & 6 students, Toronto, ON) were drawn from 16 schools that varied by neighbourhood type (pre-1946
“old/urban BE” with grid-based street layout versus post-1946 “new/inner-suburban BE” with looping street layout) and socio-economic status (low and
high SES). Physical activity was recorded by accelerometry for seven days. Only children living within 1.6 km of school were included in the analyses
(n=713; boys=339, girls=374). Generalized linear mixed models examined sex-specific differences in physical activity across four geographic
stratifications: old BE, low-SES (OL); old BE, high-SES (OH); new BE, low-SES (NL); and new BE, high-SES (NH).
Results: Children who attended schools in more affluent neighbourhoods (urban and inner-suburban) had more positive physical activity profiles.
Across school days, boys were more active in inner-suburban neighbourhoods whereas urban and inner-suburban girls’ activity levels were similar. On
the weekend, the influence of the neighbourhood environment was stronger, especially for girls and also for boys with respect to total activity and the
accumulation of moderate-to-vigorous physical activity.
Conclusion: These findings focus attention on the need to consider the broader social and temporal contexts of specific geographic locations when
planning and implementing built environment interventions to increase physical activity among children.
Key words: Accelerometer; child; built environment; physical activity
T
he built environment consists of the buildings, roads and
planned open spaces in which people live, work and perform
other daily activities (e.g., study, eat, socialize and play). The
physical layout of communities can promote or limit opportunities for physical activity. Using accelerometers to capture objective
levels of physical activity, Frank and colleagues observed that features of community design (increased land-use mix, street connectivity and residential density) were positively associated with the
accumulation of moderate-to-vigorous physical activity (MVPA)
and the achievement of physical activity guidelines.1 However, the
associations in that study were explored in adults living in the US.
While there is some evidence to support a link between the built
environment and children’s physical activity,2,3 most studies have
used self-reported measures of activity that show mixed results4 and
are known to have limited validity in children.5 Two studies6,7 used
objective measures of physical activity (accelerometry), yet these
focused on differences between urban and rural environments.
Where studied, suburban children tend to be most active,4 although
the findings are based on self-reported physical activity, and the
involvement of households with higher socio-economic status (SES)
could be confounding the results. Children from low SES households tend to have lower physical activity levels and to engage in
more sedentary activities.8-10
Given this SES–activity relationship, there is a need for studies
that separate SES from geographic features in investigations of children’s physical activity. Also, physical activity differs between the
sexes, girls being generally less active and less likely to achieve
physical activity recommendations than boys.11 Consequently, any
investigation into the relation of neighbourhood type and SES with
characteristics of physical activity should consider the possibility
that findings may be sex specific. To the authors’ knowledge, no
Canadian published study aiming to investigate the effect of neighbourhood on multiple aspects of young boys’ and girls’ physical
activity (total physical activity, activity intensity, time spent sedentary and minutes of light and MVPA; age 10 to 12 years) has
employed a sampling strategy that established sufficiently varied
built environment characteristics and SES; this gap provided the
incentive for Project BEAT (Built Environment and Active Transport).
The City of Toronto was the study site for Project BEAT. Marked differences in the built environment can be found across Toronto. The
inner-city is dominated by “older” traditional neighbourhoods (pre
World War II),12 but improved mobility options and demands for
Author Affiliations
1. School of Health and Human Performance, Dalhousie University, Halifax, NS
2. Faculty of Kinesiology and Physical Education, University of Toronto, Toronto, ON
3. School of Urban and Regional Planning, Ryerson University, Toronto, ON
4. Department of Geography, University of Toronto Mississauga, Mississauga, ON
Correspondence: Michelle R. Stone, School of Health and Human Performance,
Dalhousie University, 6230 South Street, PO Box 15000, Halifax, NS B3H 4R2,
Tel: 902-494-1167, Fax: 902-494-1084, E-mail: [email protected]
Acknowledgements: This research was funded by the Built Environment, Obesity
and Health Strategic Initiative of the Heart and Stroke Foundation and the Canadian
Institutes of Health Research.
S9
EXPLORING ACTIVITY BY NEIGHBOURHOOD TYPE AND SES
Figure 1.
Map identifying the 16 schools in the City of Toronto participating in Project BEAT, classified into four different
neighbourhood types
Kilometres
“Old” Neighbourhood: Mostly developed before 1946; primarily grid-based street layout.
“New” Neighbourhood: Mostly developed after 1946; primarily looped street layout.
affordable housing spurred a post-war suburban housing revolution.
As a result, conventional suburban neighbourhoods dominate the
inner-suburban Toronto. This part of the city also captures some of
Canada’s earliest experiments with planned urban form, such as the
Don Mills community and tower neighbourhoods.12,13 Over the last
two decades, however, pockets within some inner-city neighbourhoods have been re-developed, a trend that has been supported by
favourable policy and market conditions. With the exception of these
re-urbanized residential blocks in the inner-city and the tower neighbourhoods in the inner-suburbs, the era of development can reasonably be used as a proxy for neighbourhood types in Toronto. Within
the older central city, street networks are more commonly connected
(gridded), have a higher density of intersections and shorter straight
blocks, and include higher building densities and mixed use. In the
newer inner-suburbs, the neighbourhood streets are largely curvilinear with clear hierarchy, land uses are segregated, housing density
is lower, and there is more open space than in the older neighbourhoods.12,14 SES varies widely across these urban (older) and innersuburban (newer) settings. This is an important factor, as a household’s choices regarding opportunities for physical activity and the
safety of engaging in physical activity are also affected by level of SES.
This unique landscape supports our objective to classify neighbourhoods according to neighbourhood type and SES in an investigation of how neighbourhoods influence the physical activity
patterns of children in Toronto. This is a novel design that addresses the inherent gaps in the built environment and physical activity literature.
S10 REVUE CANADIENNE DE SANTÉ PUBLIQUE • VOL. 103, SUPPLÉMENT 3
METHODS
Experimental design
Children’s physical activity levels in the City of Toronto were examined. From January 2010 to June 2011, all elementary/intermediate
schools in the Toronto District School Board with Grade 5 and 6
students (n=469) received an invitation to participate. A pool of
interested schools was generated, and 16 schools were selected that
varied with respect to neighbourhood type and level of SES. Two
neighbourhood classifications were created on the basis of the period of neighbourhood development: urban (old BE) – older built
environment with primarily grid-based street layout – versus innersuburban (new BE) incorporating newer built environment with
primarily looped street layout (Figure 1). Neighbourhood era of
development was computed at the scale of the census dissemination area (DA). DAs are the smallest geographic units (0.18±0.39
km2) for which detailed public census data (by Statistics Canada) are
available. All DAs in which >50% of the residential building stock
was developed before 1946 were identified as urban/old neighbourhoods. The year 1946 was selected to represent a proxy for pre
and post World War II neighbourhoods. Development patterns in
Toronto changed noticeably in the post-war era as a result of a widespread implementation of the “planned neighbourhood” design
concept.12,13 For the purpose of this study, we assumed that the general physical qualities of a neighbourhood (i.e., neighbourhood
type) would be similar within a 1.6 km radius of a school location.
Children living >1.6 km from school were deemed eligible for
Table 1.
Descriptive Characteristics of Sample, by Sex (n=713)
Characteristic
OL (n=158)
Boys
Girls
Sample size
79
79
Mean age, years (SD)
10.9 (0.7)§
11.0 (0.7)§
Mean height, cm (SD)
147.1 (8.6)
146.3 (8.3)
Mean weight, kg (SD)
43.3 (12.5)
40.3 (9.8)
18.7 (3.4)
Mean body mass index, kg/m2 (SD) 19.7 (4.3)
BMI category||
Normal weight, %
65.8
75.9
Overweight or obese, %
34.2
24.1
Neighbourhood Classification
OH (n=194)
NL (n=214)
NH (n=147)
Boys
Girls
Boys
Girls
Boys
Girls
86
108
103
111
71
76
11.1 (0.6)
11.2 (0.6)
11.0 (0.6)
11.0 (0.6)§
11.2 (0.6)*
11.3 (0.6)*‡
147.2 (6.8)
149.0 (8.0)‡
147.2 (7.1)
145.0 (12.2)†§ 147.4 (7.5)
150.0 (8.0)‡
40.0 (9.0)
39.7 (7.6)
42.9 (11.3)
40.9 (10.1)
43.6 (10.8)
42.1 (9.2)
18.3 (3.1)§
17.8 (2.5)‡
19.6 (4.1)
19.1 (3.7)†
19.9 (3.9)†
18.6 (3.2)
75.6
24.4
86.1‡
13.9‡
61.2
38.8
62.2†
37.8†
62.0
38.0
73.7
26.3
* Significantly different from OL, p<0.05
† Significantly different from OH, p<0.05
‡ Significantly different from NL, p<0.05
§ Significantly different from NH, p<0.05
|| International Obesity Task Force classification15
OL=Old built environment (urban), low socio-economic status (SES); OH=Old built environment (urban), high SES; NL=New built environment (inner-suburban),
low SES; NH=New built environment (inner-suburban), high SES
school bus transportation as they were considered to reside outside
the school catchment area (www.tdsb.on.ca).
Two classifications of SES for neighbourhoods around the school
locations were also created (Low SES and High SES) according to
the median household income reported in the 2006 Population
Census of Canada. For each school (n=469), the median household
income within an 800 m (i.e., 0.8 km/0.5 mile) straight line buffer
distance was estimated by taking a median of the DA-level household incomes. Schools with the lower 50th percentile values were
identified as the Low SES schools. The SES was measured at a larger geographic scale (than neighbourhood type) in order to capture
the general socio-economic characteristics of a school’s student
population, who may live in various neighbourhoods (i.e., in different DAs) near the school. Half of the surveyed schools (i.e., eight
schools) were Low SES schools, and the other half were High SES
schools. Consent was obtained from participating school boards,
individual schools, parents and students. Student participation was
voluntary.
A total of 1,027 parents/guardians gave consent for their children to participate (boys, n=478; girls, n=549). Height and weight
measurements were taken and accelerometer-measured physical
activity data collected on a total of 1,001 children. Of those, 85.5%
had at least three weekdays and one weekend day of valid data
(n=856; boys=389, girls=467). Analyses were conducted only on
children living within 1.6 km of school (n=713; boys=339,
girls=374; mean age 11.1±0.6 years). With the use of age- and sexspecific body mass index (BMI) cut-points provided by the International Obesity Task Force,15 participants were classified as of
normal weight, overweight or obese.
Measurement of physical activity
Children’s physical activity was objectively measured for seven days
using accelerometry (ActiGraph GT1M; ActiGraph LLC, Pensacola,
FL, US). A 5 s epoch (interval) was used to capture the rapid transitions in activity typical in children.16 For inclusion in data analysis, each child required a minimum of 10 hours of wearing time for
at least 3 weekdays and 1 weekend day.17 Time spent at various levels of movement intensity was classified according to published
thresholds in children18 and used to determine levels of physical
activity during school days (weekdays; WD) and weekends (WE).
Physical activity variables of interest included total physical activity (TPA; counts.day–1), mean counts (MC; counts.min–1), time spent
sedentary (% of day) and minutes of light-intensity activity (LPA)
and moderate-to-vigorous activity (MVPA). Data collection took
place during the spring/summer (April to June) and fall (September
to December) school periods to limit any seasonal effect.
Statistical analyses
Generalized linear mixed models were used to examine sex-specific
differences in accelerometry summary measures (TPA, MC, sedentary behaviour, LPA and MVPA) for WD and WE across four neighbourhood classifications based on neighbourhood type and SES:
old BE, low-SES (OL); old BE, high-SES (OH); new BE, low-SES (NL);
new BE, high-SES (NH). Random effects at classroom levels were
included to account for possible variability (i.e., clustering of
accelerometry data among different classrooms) and adjust for any
clustering effects. Sex-specific differences in descriptive characteristics (age, height, weight, BMI and proportion of normal weight
and overweight/obese participants) were also explored across neighbourhood classifications. Estimated means were compared and significant differences tested using the Sequential Bonferroni method.
The alpha level was set at 0.05. SPSS version 19.0 was used for all
analyses.
RESULTS
General characteristics
Data for 713 participants are presented (mean age 11.1±0.6 years;
boys, n=339, girls, n=374, Table 1). For boys, only age and BMI differed among neighbourhoods (boys in urban neighbourhoods were
slightly younger [OL] and had lower BMIs [OH] than boys in NH
neighbourhoods, p<0.05). For girls, there were significant differences in age, height, BMI and weight classification. Girls in low SES
neighbourhoods were younger and shorter (particularly those in
NL neighbourhoods) than girls in high SES neighbourhoods. Furthermore, compared with girls in OH neighbourhoods, those in NL
neighbourhoods had greater BMIs, and a significantly greater proportion were classified as being overweight or obese (p<0.05, Table
1).
Weekday physical activity
The type of neighbourhood most conducive to high levels of PA
across school days differed between boys and girls. Boys in innersuburban, high SES neighbourhoods had the highest activity levels;
CANADIAN JOURNAL OF PUBLIC HEALTH • NOVEMBER/DECEMBER 2012 S11
358,785 (134,686)†
407,089 (146,110)*
487,736 (129,896)
533,073 (147,799)
172.7 (5.2)
166.5 (4.8)
169.5 (5.8)
342.0 (23.5)†§ 192.9 (3.5)
592.6 (22.8)*†‡ 470.3 (26.7)*‡ 195.1 (3.1)
452.8 (20.1)§
515.5 (20.4)§
483.1 (23.9)*‡ 181.2 (3.7)
WE
155.8 (5.5)
WD
LPA
(min)
357.6 (25.6)†§ 194.0 (3.9)
474.7 (21.9)§
MC
(counts·min–1)
WD
WE
WD
42.5 (2.1)
38.0 (1.8)
38.0 (1.8)
36.8 (2.0)
WE
28.7 (2.1)
23.9 (1.8)†
31.1 (1.9)*‡
22.0 (2.0)†
MVPA
(min)
73.9 (0.7)*†‡
78.6 (0.7)§
76.9 (0.7)§
77.3 (0.7)§
379.9 (18.2)†§ 306.8 (23.6)†
MC
(counts·min–1)
WD
WE
447.2 (18.6)*‡ 353.8 (24.1)
400,625 (109,118)‡ 303,979 (180,593)†
147.5 (4.4)
133.6 (5.2)†
168.6 (3.3)
166.5 (3.9)
WD
26.6 (1.5)
22.6 (1.3)†
29.3 (1.3)‡
24.6 (1.5)
WE
17.8 (1.9)†
16.3 (1.6)†
25.2 (1.6)*‡§
15.9 (1.8)†
MVPA
(min)
78.7 (0.7)‡
81.9 (0.6)†§
78.8 (0.6)‡
80.4 (0.7)
82.2 (0.8)†
83.3 (0.7)†
79.4 (0.7)*‡§
82.7 (0.8)†
SB
(% of day)
WD
WE
|| Presented as mean (SE).
TPA=total physical activity; MC=mean counts; LPA=light physical activity; MVPA=moderate-to-vigorous physical activity; SB=% of day spent sedentary; WD=weekdays; WE=weekend
341.7 (16.3)†§ 295.3 (21.0)†
359,446 (90,999)†§ 286,712 (92,038)†
154.8 (4.4)§
WE
149.7 (5.0)
WD
LPA
(min)
177.6 (3.8)
406,311 (110,694)‡ 375,598 (181,743)*‡§ 444.2 (15.8)*‡ 428.0 (20.5)*‡ 166.7 (3.3)
WE
288,726 (133,021)†
WD
392,254 (114,890)
TPA
(counts·day–1)
Influence of Neighbourhood Type and SES on Characteristics|| of Accelerometer-measured Physical Activity in Girls
OL: Old built environment,
low SES (n=79)
OH: Old built environment,
high SES (n=108)
NL: New built environment,
low SES (n=111)
NH: New built environment,
high SES (n=76)
Girls (n=374)
Table 3.
77.2 (0.9)*‡
81.8 (0.8)†,§
77.0 (0.8)*‡
81.3 (0.9)†§
SB
(% of day)
WD
WE
|| Presented as mean (SE).
TPA=total physical activity; MC=mean counts; LPA=light physical activity; MVPA=moderate-to-vigorous physical activity; SB=% of day spent sedentary; WD=weekdays; WE=weekend
429,351 (196,481)*‡
483,650 (127,870)
WE
335,549 (169,399)†§
WD
478,770 (135,869)
TPA
(counts·day–1)
Influence of Neighbourhood Type and SES on Characteristics|| of Accelerometer-measured Physical Activity in Boys
OL: Old built environment,
low SES (n=79)
OH: Old built environment,
high SES (n=86)
NL: New built environment,
low SES (n=103)
NH: New built environment,
high SES (n=71)
Boys (n=339)
Table 2.
the overall intensity of activity they accumulated (mean counts) was significantly greater
and they spent a significantly lower proportion of their day sedentary compared with
boys in all other neighbourhoods (p<0.05,
Table 2). However, the accumulation of LPA
and MVPA across school days was no different
from that in other neighbourhoods. There was
also a trend for WD total activity to be higher
in NH neighbourhoods than OL neighbourhoods (p=0.07). For girls, an inner-suburban,
low SES neighbourhood was least enhancing
with regard to physical activity. Compared
with those going to schools in high SES neighbourhoods, these girls spent a significantly
greater proportion of their day sedentary, and
the activity that they accumulated across the
day was less intense; they also accumulated
less total activity and, in particular, less MVPA
(p<0.05, Table 3). The overall WD activity profile of girls in urban, low SES neighbourhoods
was also less intense. Similar to boys, the accumulation of LPA on weekdays was similar
across neighbourhoods.
Weekend physical activity
For boys, WE activity profiles were strongest
among children in high SES neighbourhoods
(urban and inner-suburban, p<0.05); however, the accumulation of LPA was similar
across neighbourhoods (Table 2). Boys in high
SES, urban neighbourhoods also had greater
total activity and accumulated more MVPA
than those living in more economically disadvantaged communities. For girls, WE activity levels were highest among those situated
in urban, economically advantaged neighbourhoods: these girls accumulated significantly more total activity and MVPA, and
spent a lower proportion of their day sedentary than did girls in all other neighbourhoods
(p<0.05, Table 3). Compared with girls in low
SES neighbourhoods, the overall intensity of
their activity profile was also higher, and they
accumulated significantly more LPA over the
weekend than girls in inner-suburban, high
SES neighbourhoods.
DISCUSSION
This study aimed to investigate the relationship between school neighbourhood type
(based primarily on the period of neighbourhood development) and SES and physical
activity in children using a sampling frame
that purposely located schools in varying
neighbourhoods to ensure that there was variability in built environment characteristics
and SES. Our work generated three key lessons.
Lesson 1: Area level SES factors matter
Children who attend schools in more affluent neighbourhoods,
irrespective of neighbourhood type (urban and inner-suburban), have
more positive physical activity profiles across the week. The observation
of high physical activity levels among children in inner-suburban
high SES neighbourhoods corresponds with previous accounts from
self-reported PA data.4 Families in newer neighbourhoods with economic means may encourage structured, localized, higher-intensity
activities that compensate for potential reductions in habitual physical activity associated with design features that inhibit walking or
unstructured play. Less affluent school neighbourhoods have been
shown to have social and physical environments less conducive to
maintaining healthy weights and levels of physical activity.19,20 They
may lack recreational facilities or have facilities that require a fee.21
Less affluent neighbourhoods are also more likely to be perceived as
unsafe.19 Perceived threats to safety are one of the biggest barriers to
children’s independent play and mobility.22
Overall, this finding highlights the need for interventions
addressing inequalities at the individual and neighbourhood levels.
These may include built environment modifications, but it is likely that a broader intervention approach is required in alleviating
safety concerns, increasing social capital and cohesion, and subsidizing opportunities for physical activity.
Lesson 2: The influence of the neighbourhood
environment may vary over time
On the weekend, the combination of affluence and an urban environment becomes particularly important in raising children’s physical
activity profiles, especially for girls and also for boys with respect to total
activity and the accumulation of MVPA. When compared with children from lower SES neighbourhoods, the results take on a more practical significance. For example, the approximately 3%-4% difference in
time spent being sedentary on the weekend among groups amounts to
an extra hour of sedentary time for children in lower SES neighbourhoods; these children also accumulate seven to nine fewer minutes of
MVPA than their urban, high SES neighbourhood counterparts on the
weekend. Toronto’s urban neighbourhoods are older and have greater
street connectivity, and in more affluent areas where safety concerns
are low might provide a favourable environment for accessing opportunities and engaging in outdoor activities and play. Since children
potentially have more discretionary input into decisions over time use
during the weekend, the effect of this type of environment on physical
activity might be stronger during that time period.
Notably, this finding highlights that the relationship between neighbourhood type (and likely more broadly the “built environment”) and
physical activity is temporally heterogeneous. That is, the strength of
association between features of the built environment and physical
activity varies at different times of the day or week – for example, as the
spatial, temporal and institutional constraints (e.g., family structure,
access to daycare, location of work, employment status, access to cars)
facing households also changes over time.23 This has important implications for what and when built environment interventions might
work in increasing the physical activity of children.
Lesson 3: Gender and the type of physical activity
measured matters
The impact of neighbourhood classification on aspects of physical
activity is different for boys and girls. For example, girls in urban
and inner-suburban high SES neighbourhoods had similar weekday activity levels that were significantly higher than the levels of
girls in low SES neighbourhoods. This was not the case for boys:
those in high SES, inner-suburban neighbourhoods had significantly greater activity profiles than their urban counterparts. Girls
may be granted less independent mobility than boys,24 and this
might be further amplified in less affluent neighbourhoods because
of heightened parental concerns regarding personal safety.
Additionally, the impact of neighbourhood appeared weaker for
some characteristics of activity: for boys, the accumulation of
MVPA across the school week and the accumulation of LPA over
the weekend were similar across neighbourhoods. Older neighbourhoods, with traditionally greater street connectivity, may
encourage walking for various activities, therefore one might expect
to see a greater accumulation of LPA among children living in these
neighbourhoods. Yet our data demonstrate that for the most part,
the accumulation of LPA is similar among neighbourhoods; only
over the weekend did differences arise, in girls only, when those in
urban, high SES neighbourhoods accumulated more LPA than those
in inner-suburban, high SES neighbourhoods. Overall, these findings emphasize that built environment interventions may have
variable impact on different types of physical activity and groups
of children (e.g., boys and girls).
Strengths and limitations
The strengths of this study include the large sample, the sampling
frame and the use of an objective measure of physical activity to
examine multiple aspects of physical activity across both school
days and over the weekend. Our collection of high-frequency physical activity data was particularly appropriate for quantifying children’s activity.16,25,26
The limitations of the study include the narrow age range of children sampled and the investigation of Toronto neighbourhoods,
which do limit the generalizability of the findings to other age
groups and geographic locations. The present study did not examine the influence of micro-level community design and land-use
characteristics (e.g., connectivity, access/proximity to recreational
facilities, residential density). Moreover, since Toronto’s public
schools maintain small catchment areas, this research assumed that
the socio-economic and built environment near school and home
locations would generally be similar (1.6 km between school and
home). However, we recognize that different definitions of neighbourhood may have yielded different results (although such differences might be small14) and that the built environment near the
home location may be different than around the school within our
sample. For example, Mitra and colleagues compared the relative
influences of the home and school neighbourhoods on active
school transportation and found that the built environment near
home was more important in enabling walking among children.27
An exploration of the relationship between the objective qualities
of the neighbourhood of both the home and the school, and measures of physical activity, remains a focus of future investigation.
CONCLUSION
In conclusion, our findings highlight the value of geographic stratification based on neighbourhood type and SES in cross-sectional
analyses of accelerometry data. Our work offers three key lessons:
one, that physical activity varies more by level of school neighCANADIAN JOURNAL OF PUBLIC HEALTH • NOVEMBER/DECEMBER 2012 S13
bourhood affluence than neighbourhood type; two, that broader
relationships between the built environment and physical activity
may vary temporally; and three, that the influence of the built environment is different for boys and girls, and varies according to the
type of physical activity. In planning and implementing built environment interventions to increase physical activity among children, these lessons focus attention on the need to consider the
broader social and temporal contexts of specific geographic locations.
REFERENCES
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
Frank LD, Schmidt TL, Sallis JF, Chapman J. Linking objectively measured
physical activity with objectively measured urban form: Findings from
SMARTRAQ. Am J Prev Med 2005;28:117-25.
Davison K, Lawson CT. Do attributes in the physical environment influence
children’s physical activity? A review of the literature. Int J Behav Nutr Phys Act
2006;3:19 (doi:10.1186/1479-5868-3-19).
Rahman T, Cushing R, Jackson RJ. Contributions of built environment to
childhood obesity. Mt Sinai J Med 2011;78:49-57.
Sandercock G, Angus C, Barton J. Physical activity levels of children living in
different built environments. Prev Med 2010;50:193-98.
Pate RR, Freedson PS, Sallis JF, Taylor WC, Sirard J, Trost SG, Dowda M.
Compliance with physical activity guidelines: Prevalence in a population of
children and youth. Ann Epidemiol 2002;12:303-8.
Loucaides CA, Chedzoy SM, Bennett N. Differences in physical activity levels between urban and rural school children in Cyprus. Health Educ Res
2004;19:138-47.
Tremblay MS, Barnes JD, Copeland JL, Esliger DW. Conquering childhood
inactivity: Is the answer in the past? Med Sci Sports Exerc 2005;37:1187-94.
Drenowatz C, Eisenmann J, Pfeiffer K, Welk G, Heelan K, Gentile D, Walsh D.
Influence of socio-economic status on habitual physical activity and sedentary behavior in 8- to 11-year old children. BMC Public Health 2010;10:214.
Available at: http://www.biomedcentral.com/1471-2458/10/214 (Accessed
October 5, 2011).
Ferreira I, Horst K van der, Wendel-Vos W, Kremers S, van Lenthe FJ, Brug J.
Environmental correlates of physical activity in youth – a review and update.
Obes Rev 2007;8:129-54.
Maher CA, Olds CS. Minutes, MET-minutes, and METs: Unpacking socioeconomic gradients in physical activity in adolescents. J Epidemiol Community
Health 2011;65:160-65.
Colley R, Garriguet D, Janssen I, Craig CL, Clarke J, Tremblay MS. Physical
activity of Canadian children and youth: Accelerometer results from the 2007
to 2009 Canadian Health Measures Survey. Health Reports (Statistics Canada,
82-003) 2011;22:1-10.
Hess PM. Avenues or arterials: The struggle to change street building practices in Toronto, Canada. J Urban Design 2009;14:1-28.
Sewell J. Don Mills: Canada’s first corporate suburb. In: Sewell J, The Shape of
the City: Toronto Struggles with Modern Planning. Toronto, ON: University of
Toronto Press, 1993; chapter 3.
Mitra R, Buliung RN. Built environment correlates of active school transportation: Neighbourhood and the modifiable areal unit problem. J Transport
Geography 2012;20:51-61.
Cole TJ, Bellizzi MC, Flegal KM, Dietz WH. Establishing a standard definition
for child overweight and obesity worldwide: International survey.
BMJ 2000;320:1240.
Stone MR, Rowlands AV, Middlebrooke AR, Jawis MN, Eston RG. The pattern
of physical activity in relation to health outcomes in boys. Int J Pediatr Obes
2009;4:306-15.
Stone MR, Rowlands AV, Eston RG. Characteristics of the activity pattern in
normal weight and overweight boys. Prev Med 2009;49:205-8.
Stone MR, Rowlands AV, Eston RG. Relationships between accelerometerassessed physical activity and health in children: Impact of the activityintensity classification method. J Sports Sci Med 2009;8:136-43.
Oliver LN, Hayes MV. Neighbourhood socio-economic status and the prevalence of overweight Canadian children and youth. Can J Public Health
2005;96:415-20.
20. Veugelers P, Sithole F, Zhang S, Muhajarine N. Neighborhood characteristics
in relation to diet, physical activity and overweight of Canadian children.
Int J Pediatr Obes 2008;3:152-59.
21. Moore LV, Diez Roux AV, Evenson KR, McGinn AP, Brines SJ. Availability of
recreational resources in minority and low socioeconomic status areas. Am J
Prev Med 2008;34:16-22.
22. Veitch J, Bagley S, Ball K, Salmon J. Where do children usually play? A qualitative study of parents’ perceptions of influences on children’s active freeplay. Health & Place 2006;12:383-93.
23. Mitra R, Buliung RN, Faulkner G. Spatial clustering and temporal mobility of
walking school trips in the Greater Toronto Area, Canada. Health & Place
2010;6:646-55.
24. Page AS, Cooper AR, Griew P, Davis L, Hillsdon M. Independent mobility in
relation to weekday and weekend physical activity in children aged 10-11
years: The PEACH Project. Int J Behav Nutr Phys Act 2009 Jan 7;6:2.
25. Bailey RC, Olson J, Pepper SL, Porszasz J, Barstow TJ, Cooper DM. The level
and tempo of children’s physical activities: An observational study. Med Sci
Sports Exerc 1995;27:1033-41.
26. Stone MR, Rowlands AV, Eston RG. The use of high-frequency accelerometry
monitoring to assess and interpret children’s activity patterns. In: Jürimaë T,
Armstrong N, Jürimaë J (Eds.), Children and Exercise XXIV. London: Routledge, 2008;150-53.
27. Mitra R, Buliung RN, Roorda MJ. The built environment and school travel
mode choice in Toronto, Canada. Transportation Research Record
2010;2156:2150-59.
RÉSUMÉ
Objectif : Les occasions pour un enfant de faire de l’activité physique et
la possibilité de le faire en toute sécurité sont influencées par les éléments
du milieu bâti (MB). Notre étude porte sur la relation entre le type de
quartier, le statut socioéconomique (SSE) et l’activité; la base
d’échantillonnage utilisée contenait volontairement des écoles de divers
quartiers pour assurer la variabilité des caractéristiques du MB et du SSE.
Méthode : Les participants (1 027 élèves de 5e et de 6e année à Toronto,
en Ontario) ont été choisis dans 16 écoles qui variaient selon le type de
quartier (« MB ancien/urbain » d’avant 1946 avec rues quadrillées ou
« MB nouveau/de la proche banlieue » d’après 1946 avec rues en boucles)
et le statut socioéconomique (SSE faible ou élevé). L’activité physique a
été enregistrée par accélérométrie pendant sept jours. Seuls les enfants
vivant à moins de 1,6 km de l’école ont été inclus dans l’analyse (n=713;
garçons=339, filles=374). Au moyen de modèles linéaires mixtes
généralisés, nous avons examiné les différences par sexe dans l’activité
physique entre quatre stratifications géographiques : MB ancien, faible
SSE; MB ancien, SSE élevé; MB nouveau, faible SSE; et MB nouveau, SSE
élevé.
Résultats : Les enfants qui fréquentaient les écoles des quartiers aisés
(urbains et de la proche banlieue) avaient des profils d’activité physique
plus positifs. Les jours d’école, les garçons des quartiers de la proche
banlieue étaient plus actifs, tandis que les niveaux d’activité des filles
étaient semblables en milieu urbain et dans la proche banlieue. Les fins
de semaine, l’influence de l’environnement du quartier était plus forte,
surtout chez les filles, mais aussi chez les garçons en ce qui a trait à
l’activité totale et à l’accumulation d’activité physique modérée à
vigoureuse.
Conclusion : Ces résultats montrent qu’il faut examiner le contexte
social et temporel des lieux géographiques lorsqu’on planifie et que l’on
met en œuvre des interventions sur le milieu bâti pour accroître l’activité
physique chez les enfants.
Mots clés : accéléromètre; enfant; milieu bâti; activité physique
Linking Childhood Obesity to the Built Environment: A Multi-level
Analysis of Home and School Neighbourhood Factors Associated
With Body Mass Index
Jason A. Gilliland, PhD,1-3 Claudia Y. Rangel, MA,1 Martin A. Healy, MSc,1 Patricia Tucker, PhD,4
Janet E. Loebach, MEDes,1 Paul M. Hess, PhD,5 Meizi He, PhD,6 Jennifer D. Irwin, PhD,7 Piotr Wilk, PhD3,8
ABSTRACT
Objectives: This study examines environmental factors associated with BMI (body mass index) levels among adolescents with the aim of identifying
potential interventions for reducing childhood obesity.
Methods: Students (n=1,048) aged 10-14 years at 28 schools in London, ON, completed a survey providing information on age, sex, height, weight,
home address, etc., which was used to construct age-sex adjusted BMI z-scores. The presence of recreation opportunities, fast-food outlets and
convenience stores was assessed using four areal units around each participant’s home and school neighbourhood: “circular buffers” encompassing
territory within a straight-line distance of 500 m and 1000 m; and “network buffers” of 500 m and 1000 m measured along the street network. School
neighbourhoods were also assessed using school-specific “walksheds”. Multilevel structural equation modeling techniques were employed to
simultaneously test the effects of school-environment (Level 2) and home-environment (Level 1) predictors on BMI z-scores.
Results: Most participants (71%) had a normal BMI, 16.9% were overweight, 7.6% were obese, and 4.6% were considered underweight. Multilevel
analyses indicated that built environment characteristics around children’s homes and schools had a modest but significant effect on their BMI. The
presence of public recreation opportunities within a 500 m network distance of home was associated with lower BMI z-scores (p<0.05), and fast-food
outlets within the school walkshed was associated with higher BMI z-scores (p<0.05).
Conclusion: Interventions and policies that improve children’s access to publicly provided recreation opportunities near home and that mitigate the
concentration of fast-food outlets close to schools may be key to promoting healthy lifestyles and reducing childhood obesity.
Key words: Obesity; child; adolescent; environment; diet; recreation
C
hildhood obesity has become a critical public health issue in
Canada, as rates have tripled over the past three decades.1
Over one in four Canadian children are either overweight or
obese (17% and 9% respectively).2 The increased prevalence of
childhood obesity has been linked to the concurrent rise of physical health problems normally associated with adults, including
Type 2 diabetes, hypertension, heart disease and pulmonary diseases, as well as socio-psychological afflictions such as discrimination, behavioural problems, negative self-esteem, anxiety and
depression.3-6 A rapidly expanding avenue of research suggests that
rising rates of obesity are due not only to individual-level factors
(i.e., genetics), but also to characteristics of our local built environments that may be encouraging or discouraging the healthy
diets or active lifestyles associated with healthy body weights.7-10
Previous research has confirmed that obesity is linked to the consumption of energy-rich, fast foods.11 Large-scale US studies have
found that adult obesity rates are positively associated with the
density of neighbourhood fast-food outlets12 and convenience
stores.13 Much of the emphasis on the link between food and children’s health focuses on advertising14 or food policies within
schools;15-17 however, some policy-makers and public health professionals are shifting their focus to the food environments surrounding schools, as new research indicates that many children
visit food retailers on their way to and from school, mostly filling
up on high-sugar or high-fat, energy-dense foods.18 Several studies
have shown that fast-food outlets are more prevalent near
schools19,20 and in low-income neighbourhoods,21,22 suggesting that
these vulnerable populations may be at heightened risk of developing poor eating habits as a result of increased exposure to
unhealthy foods. Furthermore, it has been shown in London, ON,
Author Affiliations
1. Human Environments Analysis Laboratory, Department of Geography, University
of Western Ontario, London, ON
2. School of Health Studies and Department of Paediatrics, University of Western
Ontario, London, ON
3. Scientist, Children’s Health Research Institute & Lawson Health Research Institute,
London, ON
4. School of Occupational Therapy, University of Western Ontario, London, ON
5. Department of Geography & Program in Planning, University of Toronto, Toronto, ON
6. Department of Health & Kinesiology, The University of Texas at San Antonio,
San Antonio, TX
7. School of Health Studies, University of Western Ontario, London, ON
8. Department of Paediatrics and Department of Epidemiology & Biostatistics,
University of Western Ontario, London, ON
Correspondence: Jason A. Gilliland, Dept. of Geography, University of Western
Ontario, 1151 Richmond St, London, ON N6A 5C2, Tel: 519-661-2111, ext. 81239,
Fax: 519-661-3750; E-mail: [email protected]
Acknowledgements: This study was supported by research grants from the Heart
and Stroke Foundation of Canada, the Green Shield Canada Foundation and the
Canadian Institutes of Health Research’s Institutes of Human Development, Child and
Youth Health, and Nutrition, Metabolism and Diabetes.
CHILD OBESITY AND THE BUILT ENVIRONMENT
Figure 1.
Illustration of different areal units used for characterizing neighbourhood environments
that the presence of fast-food outlets and convenience stores within 1 km of school is linked to increased junk food purchasing23 and
poorer diets.24 However, a national study of Canadian students in
grades 6-10 found that increased number of food retailers within
1 km of the school did not increase the likelihood of the students
being overweight.25 More research is needed to confirm the links
between the local food environment and childhood obesity.
Increased physical activity is associated with reduced obesity and
other health benefits for children and youth.26 Unfortunately,
Canadian children devote 62% of their free time (after school and
weekends) to sedentary pursuits, and activity levels decline with
age.27 A growing body of research suggests that the layout and
design of children’s neighbourhood environments may be a key
facilitator or barrier to physical activity.28 Access to opportunities
for physical activity, such as public parks and recreation facilities,
has been repeatedly associated with higher rates of physical activity among children and adolescents.8,28-30 Opportunities located
within walking distance of home may be doubly important for
stimulating active behaviours, as both the route and the destination
contribute to overall activity levels.31 It has been argued that children from low-income households tend to have fewer opportunities for health-promoting activity because of a reduced ability to
afford fee-based recreation programs as well as systemic sociospatial inequities with respect to public resources.32 However, previous research in London has shown that publicly provided recreation opportunities are equitably distributed with no obvious
socio-spatial disparities.33
The purpose of this study is to identify built environment factors associated with high BMI levels among adolescents in London,
ON, to help identify potential environmental interventions for
reducing childhood obesity.
METHODS
Survey and outcome measure
Students in grades 6, 7 and 8 at 28 elementary schools within the
city of London, ON, were invited to complete a questionnaire that
asked for their home postal code, sex, age, height, weight and various health-related questions. The sampled schools were selected
from neighbourhoods of varying built environments across the city
(see Figure 1) to represent the full diversity of environmental factors that children experience around their homes and schools. Prior
to school recruitment, ethics approval was obtained from University of Western Ontario’s Research Ethics Board and the ethics
boards at both the Thames Valley District School Board and the
London District Catholic School Board. Informed written consent
was obtained from 1,048 adolescents and their parents before data
collection. A total of 966 out of 1,048 children aged 10-14 years
who participated in the study provided the complete information
on age, sex, height and weight - information needed to construct
their age- and sex-adjusted BMI z-score - and 891 of those provided accurate address information for deriving measures of their
home built environment.
Self-reported height and weight were used to estimate the body
mass index (BMI) of participants by dividing weight in kilograms by
height in metres squared; BMI z-scores were calculated to control
for differences by age and sex. Following procedures established in
previous studies,34 we calculated age- and sex-specific BMI z-scores
based on the World Health Organization growth curves, which in
turn are based on large samples of children selected to represent
optimal growth.35
ured along the street network. To assess the built environment
around the school, we used the same four circular and network
buffers at 500 m and 1000 m from the school address, as well as an
additional areal unit: the school “walkshed”. The walkshed is the
territory within a school’s observed catchment area that encompasses only those students living within walking distance, as
defined by the respective school boards (see Figure 1). School walksheds were generated for each school by: mapping the home postal
code of every registered student not eligible for bussing according
to school board data; selecting postal code centroids within the
school board-mandated 1600 m walking distance of the specific
school; and then merging all neighbouring city blocks that contained a selected postal code into a single walkshed polygon. We
argue that this definition of unique walksheds based on the residential locations of students within walking distance of each school
is a better representation of the local school neighbourhood than
the standard buffers commonly used by researchers.
Previously validated databases of every fast-food outlet and convenience store in the city and surrounding county were provided
by the Middlesex-London Health Unit, which is mandated to keep
a current inventory of all food retailers for the purpose of licensing
and annual health inspections. Using a master address database
provided by the City of London, every food retailer was geocoded
to its correct building. To “ground truth” the database, trained
research assistants performed on-site environmental audits within
a 1000 m buffer around six of the sample schools during the same
period as the surveys and confirmed 100% accuracy of the database: all food premises listed in the health unit inventory for those
test neighbourhoods were still in business and no new food retailers were found. Fast-food outlets were defined as restaurants with
food ordered at a counter and paid for in advance. Convenience
stores were classified as small food retailers with a floor area of less
than 1000 m. Data on school locations and public recreation
opportunities (including parks, playgrounds, arenas and recreation
centres, and sports fields/facilities) were obtained from the City of
London Planning & Development Division and had been previously validated using orthoimagery with a 30 cm ground pixel resolution.33,37 These data were used to calculate accessibility measures
for each participant using GIS (geographic information systems),
including the number of public recreation opportunities, fast-food
outlets and convenience stores within all defined home and school
neighbourhoods.
GIS analysis of environmental variables
Using ArcGIS 10.0 (Environmental Systems Research Institute Inc,
Redlands, CA), survey data for participants were geocoded to the
geographic centre of their home postal code. University of Western Ontario’s Research Ethics Board did not allow us to collect the
full address of students. Nevertheless, previous research indicates
that Canadian postal codes are suitable proxies of home neighbourhoods in urban and suburban environments.36,37 Since there is
no agreed measure of “neighbourhood”, we tested four different
areal definitions around each participant’s home postal code,
including: circular “buffers” encompassing the territory within a
distance of 1) 500 m and 2) 1000 m from the postal code centroid;
and “network buffers”, which encompass the territory reachable
within distances of 1) 500 m and 2) 1000 m from home, as meas-
Attribute tables containing the built environment variables by
school and by home postal code for each participant were linked to
questionnaire data on each student within ArcGIS 10.0 and exported for statistical analysis. The model was tested using multilevel
structural equation techniques for complex survey data. This technique allows for simultaneous testing of the effects of schoolenvironment (Level 2) and home-environment (Level 1) predictors
on a child’s BMI z-scores. The constructs representing built environment – presence of public recreation opportunities, presence of
fast-food outlets, and presence of convenience stores – were operationalized at both levels. The two sets of measurement instruments
are independent, as they refer to different geographic environments
and different units of analysis. The model was estimated using
Mplus 6.0 program (Muthén and Muthén).38
Table 1.
Demographic Characteristics of Study Participants
All
All
Sex
Boys
Girls
Age, years
10
11
12
13
14
Table 2.
Underweight
n
%
44
4.6
n
686
%
71.0
Overweight
n
%
163
16.9
n
73
%
7.6
47.4
52.6
16
28
3.5
5.5
285
401
62.2
78.9
112
51
24.4
10.0
45
28
9.8
5.5
6.8
10.9
33.0
43.3
6.0
3
9
14
17
1
4.6
8.6
4.4
4.1
1.7
44
72
224
305
41
66.7
68.6
70.2
73.0
70.7
10
10
54
78
11
15.2
9.5
16.9
18.7
19.0
9
14
27
18
5
13.6
13.3
8.5
4.3
8.6
n
966
%
100.0
458
508
66
105
319
418
58
Normal
Obese
Built Environment Characteristics of Study Participants
Circular Buffer
500 m
1000 m
n
%
n
%
Home Neighbourhood Environment
Number of public recreation opportunities
0
1
2 or more
Number of fast-food outlets
0
1
2 or more
Number of convenience stores
0
1
2 or more
School Neighbourhood Environment
0
1
2 or more
0
1
2 or more
0
1
2 or more
Network Buffer
500 m
1000 m
n
%
n
%
School Walkshed
n
%
291
147
505
30.9
15.6
53.6
89
57
797
9.4
6.0
84.5
634
175
134
67.2
18.6
14.2
252
164
527
26.7
17.4
55.9
505
104
334
53.6
11.0
35.4
181
66
696
19.2
7.0
73.8
694
92
157
73.6
9.8
16.7
378
117
448
40.1
12.4
47.5
435
134
374
46.1
14.2
39.7
177
55
711
18.8
5.8
75.4
652
122
169
69.1
12.9
18.0
344
108
491
36.5
11.5
52.1
125
103
738
12.9
10.7
76.4
77
20
869
8.0
2.1
90.0
277
307
382
28.7
31.8
39.6
185
71
710
19.2
7.4
73.5
38
47
881
3.9
4.9
91.2
471
49
446
48.8
5.1
46.2
84
58
824
8.7
6.0
85.3
682
15
269
70.6
1.6
27.9
271
20
675
28.1
2.1
69.9
54
118
794
5.6
12.2
82.2
304
254
408
31.5
26.3
42.2
20
162
784
2.1
16.8
81.2
473
281
212
49.0
29.1
22.0
187
104
675
19.4
10.8
69.9
108
27
831
11.2
2.8
86.0
RESULTS
Nearly three quarters (71.0%) of participants were categorized as
having a normal BMI, 16.9% were overweight, 7.6% were obese,
and 4.6% were considered underweight (Table 1). Boys were much
more likely to be overweight or obese than girls; however, BMI did
not vary greatly according to age.
Table 2 provides descriptive statistics for the environmental variables used in this study. It presents the number and percentage of
participants who have 0, 1, or 2 or more of the selected environmental features nearby, depending on which method is used to
define home and school neighbourhoods. Two key findings are
clear from the results presented: 1) a large percentage of children
have at least one public recreation opportunity, convenience store
and fast-food restaurant within a short walk of their home and
school; and 2) the way in which neighbourhoods are delineated, in
terms of distance from home or school and how distance is measured, has a major influence on whether the selected environmental factors appear to be accessible or not.
We employed univariate regression to determine which of the
built environment variables to include in the models on the basis
of statistically significant associations with BMI z-scores (Table 3).
The final multi-level models included the following variables for
the home environment: presence of recreation opportunities, presence of fast-food restaurants, and presence of convenience stores,
all within the 500 m network distance of home. For the school
environment, the final models included: presence of recreation
opportunities, presence of fast-food restaurants, and presence of
convenience stores, all within the school-specific walkshed. Table 4
displays the results of the multi-level analysis of the influence of the
school and home built environment on children’s BMI z-score.
Prior to testing the hypothesized model, we examined the betweenschool variability in BMI z-scores. The interclass correlation coefficient of 0.039 (p<0.05) indicated that there were statistically
significant differences in BMI z-scores across school neighbourhoods. As predicted, the results from the multi-level models show
that the home-environment predictor “presence of public recreation opportunities within 500 m network distance” had a significant negative (i.e., reducing) effect on BMI z-scores (-0.203; p<0.05).
The indicators for “presence of fast-food outlets” and “presence of
convenience stores” in the home environment, however, had no
significant effect on the outcome variable (0.012 and 0.190, respectively; p>0.05). The effect of only one of the school-environment
(Level 2) predictors, presence of fast-food outlets within the school
walkshed, was statistically significant (0.073; p<0.05), after controlling for home-environment variables.
DISCUSSION
This study of children aged 10-14 years in London, ON, found that
nearly three out of four participants (71.0%) were categorized as
having a normal BMI, 16.9% were overweight, 7.6% were obese,
Table 3.
Results of the Univariate Multi-level Regression Analyses Examining the Relationship Between School and Home Built
Environment and Children’s BMI Z-Scores
Home Neighbourhood Environment
School Neighbourhood Environment
Circular Buffer
500 m
1000 m
Estimate (SE)
Estimate (SE)
Network Buffer
500 m
1000 m
Estimate (SE)
Estimate (SE)
-0.109 (0.10)
0.204* (0.09)
0.076 (0.10)
0.043 (0.03)
0.032 (0.02)
0.044* (0.02)
-0.182* (0.09)
0.139 (0.10)
0.219* (0.10)
0.017 (0.03)
0.026 (0.02)
0.026 (0.02)
-0.136 (0.26)
0.028 (0.13)
0.138 (0.14)
-0.097* (0.04)
0.044 (0.02)
0.048* (0.02)
0.305 (0.25)
0.166 (0.26)
0.18 (0.13)
0.389 (0.24)
-0.005 (0.02)
0.021 (0.02)
School
Walkshed
Estimate (SE)
-0.017 (0.05)
0.095* (0.03)
0.057* (0.02)
* Significant at p=0.05
Table 4.
Results of the Multivariate Multi-level Regression Analysis Assessing the Influence of the School and Home Built
Environment on Children’s BMI Z-Scores
Fixed Effect
Home environment (500 m network buffer)
Recreation opportunities
Fast-food outlets
Convenience stores
School environment (walkshed)
Recreation opportunities
Fast-food outlets
Convenience stores
Random effects
Child-level residual variance
School-level residual variance
Estimate
SE
Est./SE
p-value
-0.203
0.012
0.190
0.093
0.121
0.122
-2.183
0.099
1.559
0.03
0.92
0.12
-0.019
0.073
0.020
0.041
0.034
0.021
-0.471
2.160
0.947
0.64
0.03
0.34
Estimate
SE
Est./SE
p-value
1.673
0.021
0.080
0.019
20.821
1.074
0.00
0.28
Intraclass correlation coefficient for BMI z-score=0.039 (p=0.041).
Number of children=891; number of clusters=28.
and 4.6% were considered underweight. These findings are very
similar to reported rates of overweight (17%) and obesity (9%)
among children aged 6-14 years across Canada.2 Also consistent
with previous Canadian studies2,39 is the finding that boys were
more likely to be overweight than girls; but, unlike previous studies,39 we did not see any discernible trend in BMI by age within the
limited age range of our sample.
This study makes an important empirical contribution to knowledge about the determinants of childhood obesity in Canada, as
the results of multi-level statistical analyses indicate that characteristics of the built environment around children’s homes and
schools have a modest but significant effect on their BMI. As
expected from previous studies, the presence of recreation opportunities in the home neighbourhood had a significant effect on
BMI z-scores: children who had at least one public recreation
opportunity within a 500 m walk of their home were likely to have
lower BMI z-scores than their counterparts without a recreation
opportunity nearby. Presumably, children and youth are more likely to use parks and recreation facilities if they are located within
close walking distance of their home, and lower observed BMIs may
therefore be due to increased physical activity levels associated with
greater accessibility. Indeed, this finding follows previous research
on children 11-13 year olds in London, which revealed that those
who have two or more public recreation facilities within 500 m of
their home engaged in 16.5 more minutes of physical activity per
day than children with fewer facilities in their home neighbourhood.28 On the other hand, the presence of public recreation opportunities in the school neighbourhood did not have a significant
effect on children’s BMI z-scores. This finding fits with qualitative
studies based on interviews and/or focus groups with children that
suggest they are likely to spend more of their free time, especially
on weekends, playing in the neighbourhood around the home
rather than around the school;40,41 moreover, during school hours,
children are likely to play on the school grounds rather than leave
the campus to play.
Easy access to retailers of “junk food”, such as fast-food restaurants and convenience stores around children’s homes and schools,
appeared to contribute to higher obesity levels; however, the presence of fast-food restaurants within the school walkshed was the
only food environment variable that had a statistically significant
association with higher BMI z-scores among our sample children.
This finding follows recent research on grade 7-8 students in
London, which revealed that the presence of fast-food outlets within
1 km of school is linked to increased purchasing of junk food by
students23 and poorer-quality diets (i.e., lower “healthy eating
index” scores).24 The findings are also supported by qualitative
research indicating that children who purchase junk food are likely to do it near the school, during the lunch break and the journey
to and from school, whether or not a parent may be present to evaluate their purchase.40,41 On the other hand, a study of grade 6-10
students in 178 schools across Canada did not find any statistically significant relation between the characteristics of the local food
environment around each school and the likelihood of children
being overweight; however, the study did not account for the presence/absence of recreation facilities or other opportunities for physical activity in the local environment.25
This paper has a number of methodological strengths and weaknesses that are worth noting. It makes a methodological contribution by applying multiple areal units of different shapes and sizes
(i.e., circular buffers, network buffers and school-specific walksheds)
to categorize the home and school environments of children. This
approach recognizes and attempts to account for the modifiable
areal unit problem that is inherent in most environment and
health studies.42 The school-specific walkshed is a particular innovation of this study that also proved to be the most significant
method for delineating school neighbourhoods. Another methodological improvement is that the geodatabases of environmental
variables used were locally generated by municipal organizations
and validated through rigorous processes to achieve ground-truth.
Most previous studies of this kind have relied on commercial databases, which are often incomplete and spatially inaccurate,43,44 to
identify the locations of food retailers and recreation opportunities.39,45,46
Although BMI is a widely accepted measure for comparing body
weight status at the population level, there are limitations to using
BMI for identifying overweight/obesity among children. We used
age- and sex-adjusted BMI z-scores to determine differences by age
and sex, but accuracy can be affected by factors such as ethnicity,
frame size, level of physical fitness and biologic maturation.47
Ideally, any assessment of body weight status should be calculated
using direct measurements, as self-reported heights and weights for
youth tend to underestimate the prevalence of overweight and obesity.48 Nevertheless, as reported above, rates of overweight and obesity in our sample population were not dramatically different from
rates reported for children in a recent nationwide study based on
directly measured BMI.2 Self-report measures are much less intrusive
and more feasible to collect for large samples, but as an absolute
value, the BMI should be interpreted with caution.
Another potential criticism of this research might be directed at
the fact that, because of University of Western Ontario’s Research
Ethics Board directives, we used children’s home postal codes as
the centre of their neighbourhoods rather than their actual
dwelling. Although previous research has shown that this is a common strategy in epidemiological studies and that postal codes are
adequate proxies for addresses in Canadian cities,36,37 the potential
limitations of this approach must be acknowledged: if the positional discrepancy between exact location and proxy location is
large for a given case, it may lead to the misclassification of the
presence or absence of certain environmental features.36,37 Positional discrepancy is not a significant problem in this study, however, as it has been estimated that the majority of residential
dwellings in urban and suburban London are located within 100 m
of their respective postal code centroid.37,49
It could be argued that another limitation of the study is that it
does not examine the full range of environmental variables that
have appeared in previous literature. We limited our model to the
selected variables for two reasons: 1) previous research in the same
city with children of the same age group indicated that these were
the most significant predictors of physical activity levels, junk food
purchasing and dietary quality; and 2) the statistical test for schoollevel effects revealed that the school-level built environment
accounted for only a small percentage of the variance, and therefore adding more variables would not be efficient.
This is one of the first Canadian studies to empirically establish
a relation between neighbourhood environmental factors and children’s BMI. It is also one of the only studies of its kind to focus on
a typical mid-sized North American city, as the small but growing
literature on the environment-obesity link is still dominated by
studies set in larger cities. While causal relations cannot be inferred
from these cross-sectional data and the results are not necessarily
generalizable, the study has potentially important implications for
planners, school board officials and other decision-makers involved
in the construction and management of children’s environments.
Interventions, policies and programs that increase children’s access
to high-quality, publicly provided recreation opportunities within
a short walk of home may be a key to promoting active lifestyles
and reducing obesity levels among children and youth. In addition, the study highlights the need for municipalities to consider
bylaws and policies aimed at regulating the concentration of fastfood outlets close to schools, where children are heavily exposed,
and to create incentives that encourage more healthy food options
on local menus. Given the problems associated with rising childhood obesity rates, it is imperative that further research be conducted into how environmental factors influence physical activity
levels and dietary habits among children and youth, particularly if
we are to develop interventions that promote lifelong healthy
behaviours.
REFERENCES
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
Analytic Studies and Reports. Nutrition: Findings from the Canadian Community Health Survey, Issue 1. Measured Obesity: Overweight Canadian Children and Adolescents. Ottawa, ON: Statistics Canada, 2005.
Tremblay MS, Shields M, Laviolette M, Craig CL, Janssen I, Grober SC. Fitness of Canadian children and youth: Results from the 2007-2009 Canadian
Health Measures Survey. Health Rep 2010;21:1-14.
Ball G, McCargar L. Childhood obesity in Canada: A review of prevalence
estimates and risk factors for cardiovascular diseases and type 2 diabetes.
Can J Appl Physiol 2003;28:117-40.
Figueroa-Munoz JI, Chinn S, Rona RJ. Association between obesity and asthma in 4-11 year old children in the UK. Thorax 2001;56:133-37.
Williams J, Wake M, Hesketh K, Maher F, Waters E. Health-related quality of
life of overweight and obese children. JAMA 2005;293:70-76.
Sjöberg RL, Nilsson KW, Leppert J. Obesity, shame, and depression in schoolaged children: A population-based study. Pediatrics 2005;116:e389-e392.
Frumkin H, Frank L, Jackson R. Urban Sprawl and Public Health: Designing, Planning, and Building for Healthy Communities. Washington, DC: Island Press, 2004.
Popkin BM, Duffey K, Gordon-Larsen P. Environmental influences on food
choice, physical activity and energy balance. Physiol Behav 2005;86:603-13.
Giles-Corti B, Timperio A, Bull F, Pikora T. Understanding physical activity
environmental correlates: Increased specificity for ecological models.
Exerc Sport Sci Rev 2005;33:175-81.
Leatherdale S, Papadakis S. A multi-level examination of the association
between older social models in the school environment and overweight and
obesity among younger students. J Youth Adolesc 2011;40:361-72.
Binkley JK, Eales J, Jekanowski M. The relation between dietary change and
rising US obesity. Int J Obes Relat Metab Disord 2000;24:1032-39.
Maddock J. The relationship between obesity and the prevalence of fast food
restaurants: State-level analysis. Am J Health Promot 2004;19:137-43.
Morland K, Diez Roux AV, Wing S. Supermarkets, other food stores, and obesity: The Atherosclerosis Risk in Communities Study. Am J Prev Med
2006;30:333-39.
Harris JL, Pomeranz JL, Lobstein T, Brownell KD. A crisis in the marketplace:
How food marketing contributes to childhood obesity and what can be done.
Annu Rev Public Health 2009;30:211-25.
Datar A, Nicosia N. Junk Food in Schools and Childhood Obesity: Much Ado
About Nothing? Working Paper. RAND Labor and Population, March 2009.
Fox MK, Hedley Dodd A, Wilson A, Gleason PM. Association between school
food environment and practices and body mass index of US public school
children. J Am Diet Assoc 2009;109(2):S108-S117.
Utter J, Scragg R, Percival T, Beaglehole R. School is back in New Zealand—and
so is the junk food. N Z Med J 2009;122(1290):5-8.
Cannuscio CC, Weiss EE, Asch DA. The contribution of urban foodways to
health disparities. J Urban Health 2010;87(3):381-93.
Zenk SN, Powell LM. US secondary schools and food outlets. Health & Place
2008;14:336-46.
Gilliland J. The built environment and obesity: Trimming waistlines through
neighbourhood design. In: Bunting T, Filion P (Eds.), Canadian Cities in Transition, 4th Ed. Toronto, ON: Oxford University Press, 2010.
Reidpath DD, Burns C, Garrard J, Mahoney M, Townsend M. An ecological
study of the relationship between social and environmental determinants of
obesity. Health & Place 2002;8:141-45.
Smoyer-Tomic KE, Spence JC, Raine KD, Amrhein C, Cameron N, Yasenovskiy V,
et al. The association between neighbourhood socioeconomic status and
exposure to supermarkets and fast food outlets. Health & Place 2008;14:740-54.
23. He M, Tucker P, Gilliland J, Irwin JD, Larsen K, Hess P. The influence of local
food environments on adolescents’ food purchasing behaviors. Int J Environ
Public Health 2011;8:doi:10.3390/ijperh812####.
24. He M, Tucker P, Irwin JD, Gilliland J, Larsen K, Hess P. Obesogenic neighbourhoods: The impact of neighbourhood restaurants and convenience stores
on adolescents’ food consumption behaviours. Public Health Nutr 2012;
doi:10.1017/S1368980012000584.
25. Seliske LM, Pickett W, Boyce WF, Janssen I. Association between the food
retail environment surrounding schools and overweight in Canadian youth.
Public Health Nutr 2009;12:1384-91.
26. Janssen I, LeBlanc AG. Systematic review of the health benefits of physical
activity and fitness in school-aged children and youth. Int J Behav Nutr Phys
Act 2010;7:40.
27. Colley RC, Garriguet D, Janssen I, Craig CL, Clarke J, Tremblay MS. Physical
to 2009 Canadian Health Measures Survey. Health Rep 2011;22:1-10.
28. Tucker P, Irwin J, Gilliland J, Larsen K, He M, Hess P. Environmental influences on physical activity levels in youth. Health & Place 2009;15:357-63.
29. Huston SL, Evenson KR, Bors P, Gizlice Z. Neighborhood environment, access
to places for activity, and leisure-time physical activity in a diverse North
Carolina population. Am J Health Promot 2003;18:58-69.
30. Hume C, Salmon J, Ball K. Children’s perceptions of their home and neighborhood environments, and their association with objectively measured
physical activity: A qualitative and quantitative study. Health Educ Res
2005;20:1-13.
31. Frank L, Kerr J, Chapman J, Sallis JF. Urban form relationships with walk trip
frequency and distance among youth. Am J Health Promot 2007;21:305-11.
32. Wolch J, Wilson JP, Fehrenbach J. Parks and park funding in Los Angeles: An
equity-mapping analysis. Urban Geogr 2005;26:4-35.
33. Gilliland J, Holmes M, Irwin JD, Tucker P. Environmental equity is child’s
play: Mapping public provision of recreation opportunities in urban neighbourhoods. Vulnerable Children & Youth Studies 2006;1:256-68.
34. Tremblay MS, Katzmarzyk PT, Willms JD. Temporal trends in overweight and
obesity in Canada, 1981-1996. Int J Obes 2002;26(4):538-43.
35. World Health Organization. Growth Reference Data for 5-19 years, 2007.
Available at: http://www.who.int/growthref/en/ (Accessed December 10,
2011).
36. Bow CJD, Waters NM, Faris PD, Seidel JE, Galbraith PD, Knudtson ML, et al.
Accuracy of city postal code coordinates as a proxy for location of residence.
Int J Health Geogr 2004;3:5.
37. Healy M, Gilliland J. Quantifying the magnitude of environmental exposure
misclassification when using imprecise address proxies in public health
research.
Spatial
and
Spatio-temporal
Epidemiology
2012;
doi:10.1016/j.sste.2012.02.006.
38. Muthén LK, Muthén BO. Mplus User’s Guide, Sixth Edition. Los Angeles, CA:
Muthén & Muthén, 2010.
39. Leatherdale ST, Pouliou T, Church D, Hobin E. The association between overweight and opportunity structures in the built environment: A multi-level
analysis among elementary school youth in the PLAY-ON study. Int J Public
Health 2011;56:237-46.
40. Loebach J, Gilliland J. Child guides: Exploring the use of child-led tours to
uncover perceptions and use of neighbourhood environments. Children, Youth
and Environments 2010;20(1):52-90.
41. Tucker P, Irwin JD, Gilliland, J, He M. Adolescents’ perspective of home,
school and neighborhood environmental influences on physical activity and
dietary behaviors. Children, Youth and Environments 2008;18(2):12-35.
42. Mitra R, Buliung R. Built environment correlates of active school transportation: Neighborhood and the modifiable areal unit problem. J Transport Geogr
2012;20:57-58.
43. Duvalla CS, Howard PH, Goldsberry K. Apples and oranges? Classifying food
retailers in a midwestern US city based on the availability of fresh produce.
J Hunger & Environ Nutr 2010;5(4):526-41.
44. Liese AD, Colabianchi N, Lamichhane AP, Barnes TL, Hibbert JD, Porter DE,
et al. Validation of 3 food outlet databases: Completeness and geospatial accuracy in rural and urban food environments. Am J Epidemiol 2010;172:1324-33.
45. Jeffery RW, Baxter J, McGuire M, Linde J. Are fast food restaurants an environmental risk factor for obesity? Int J Behav Nutr Phys Act 2006;3:2.
46. Alter DA, Eny K. The relationship between the supply of fast-food chains and
cardiovascular outcomes. Can J Public Health 2005;96(3):173-77.
47. Reilly JJ, Dorosty AR, Emmett PM. Identification of the obese child: Adequacy of the body mass index for clinical practice and epidemiology. Int J Obes
2000;24(12):1623-27.
48. Scholtens S, Brunekreef B, Visscher T, Smit HA, Kerkhof M, de Jongste JC,
et al. Reported versus measured body weight and height of 4-year-old children
and the prevalence of overweight. Eur J Public Health 2007;17(4):369-74.
49. Larsen K, Gilliland J, Hess P, Tucker P, Irwin J, He M. The influence of the
physical environment and sociodemographic characteristics on children’s
mode of travel to and from school. Am J Public Health 2009;99(3):520-26.
RÉSUMÉ
Objectifs : Notre étude porte sur les facteurs environnementaux associés
à l’IMC (indice de masse corporelle) d’adolescents en vue de cerner des
interventions possibles pour réduire l’obésité infantile.
Méthode : Des élèves (n=1 048) de 10 à 14 ans fréquentant 28 écoles
de London, en Ontario, ont rempli un questionnaire sur leur âge, leur
sexe, leur taille, leur poids, leur adresse personnelle, etc., lequel a servi à
construire des écarts Z ajustés selon l’âge et le sexe pour l’IMC. La
présence de possibilités de loisir, d’établissements de restauration rapide
et de dépanneurs a été évaluée à l’aide de quatre unités de surface autour
du domicile et du quartier scolaire de chaque participant : des « zones
tampons circulaires » englobant le territoire sur une distance entre 500 et
1000 m en ligne droite; et des « zones tampon de réseau » de 500 à
1000 m mesurées le long du réseau des rues. Les quartiers scolaires ont
aussi été évalués à l’aide des « bassins de marche » propres à l’école. Une
modélisation multiniveaux par équations structurelles a servi à évaluer
simultanément les effets de prédicteurs des écarts Z de l’IMC liés à
l’environnement scolaire (niveau 2) et au milieu de vie (niveau 1).
Résultats : La plupart des participants (71 %) avaient un IMC normal,
16,9 % étaient en surpoids, 7,6 % étaient obèses, et 4,6 % étaient
considérés comme étant de poids insuffisant. Des analyses multiniveaux
ont montré que les caractéristiques du milieu bâti autour du domicile et
de l’école des enfants avaient un effet mineur mais significatif sur leur
IMC. La présence d’installations de loisir publiques dans un réseau de
500 m du domicile était associée à des écarts Z d’IMC inférieurs
(p<0,05), et la présence d’établissements de restauration rapide dans le
bassin de marche de l’école était associée à des écarts Z d’IMC supérieurs
(p<0,05).
Conclusion : Les interventions et les politiques qui améliorent l’accès
des enfants à des installations de loisir publiques près de chez eux et qui
atténuent la concentration des établissements de restauration rapide près
des écoles pourraient être la clé du succès pour promouvoir les modes de
vie sains et réduire l’obésité infantile.
Mots clés : obésité; enfant; adolescent; environnement; régime
alimentaire; loisir
Smart Cities, Healthy Kids: The Association Between Neighbourhood
Design and Children’s Physical Activity and Time Spent Sedentary
Dale W. Esliger, PhD,1 Lauren B. Sherar, PhD,1 Nazeem Muhajarine, PhD2
ABSTRACT
Objectives: To determine whether, and to what extent, a relation exists between neighbourhood design and children’s physical activity and sedentary
behaviours in Saskatoon.
Methods: Three neighbourhood designs were assessed: 1) core neighbourhoods developed before 1930 that follow a grid pattern, 2) fractured-grid
pattern neighbourhoods that were developed between the 1930s and mid-1960s, and 3) curvilinear-pattern neighbourhoods that were developed
between the mid-1960s through to 1998. Children aged 10-14 years (N=455; mean age 11.7 years), grouped by the neighbourhoods they resided in,
had their physical activity and sedentary behaviour objectively measured by accelerometry for 7 days. ANCOVA and MANCOVA (multivariate analysis of
covariance) models were used to assess group differences (p<0.05).
Results: Group differences were apparent on weekdays but not on weekend days. When age, sex and family income had been controlled for, children
living in fractured-grid neighbourhoods had, on average, 83 and 55 fewer accelerometer counts per minute on weekdays than the children in the core
and curvilinear-pattern neighbourhoods, respectively. Further analyses showed that the children in the fractured-grid neighbourhoods accumulated
15 and 9 fewer minutes of moderate-to-vigorous physical activity per day and had a greater time spent in sedentary behaviour (23 and 17 minutes)
than those in core and curvilinear-pattern neighbourhoods, respectively.
Conclusion: These data suggest that in Saskatoon there is a relation between neighbourhood design and children’s physical activity and sedentary
behaviours. Further work is needed to tease out which features of the built environments have the greatest impact on these important lifestyle
behaviours. This information, offered in the context of ongoing development of neighbourhoods, as we see in Saskatoon, is critical to an evidenceinformed approach to urban development and planning.
Key words: Urban; built environment; accelerometer; lifestyle; city planning
T
he increased concern over escalating levels of chronic disease
and the emergence of the smart growth movement has yielded a series of studies investigating how aspects of the built
environment influence the physical activity and sedentary behaviours of children.1-6 Studies show that improving the built environment to “make the healthy choice the easy choice” is essential
to increasing children’s physical activity levels.7-10 In a recent systematic review by Durand et al.,11 five smart growth factors (diverse
housing types, mixed land use, housing density, compact development patterns and levels of open space) were linked with increased
levels of physical activity in children.
The burgeoning field of research related to physical activity and
the built environment is starting to influence the work of city planners. For example, the City of Saskatoon, City Centre Plan Phase 1
(2011),12 provides a policy framework that defines, directs and evaluates development of the city centre to ensure that it balances the
environmental, social and economic needs of the community.
Interestingly, since Saskatoon was incorporated as a city in 1906, its
neighbourhoods have gradually developed into what today are
three distinct design types: core neighbourhoods, fractured-grid
pattern neighbourhoods and curvilinear-pattern neighbourhoods. The core neighbourhoods, developed before 1930, represent
the oldest of the neighbourhood designs in Saskatoon, wherein the
road networks follow a grid pattern. These are typified by higher
density (11.75 people/acre), mixed-use neighbourhoods connected by straight, intersecting streets. Surrounding the core neighbourhoods are the fractured-grid pattern neighbourhoods, or
semi-suburban districts, which were developed between the 1930s
and mid-1960s. The fractured-grid pattern neighbourhoods tend to
be of lower density (9.25 people/acre), predominantly residential
and increasingly car-oriented as they are located further away from
the core. On the periphery are newer, curvilinear-pattern neighbourhoods that were developed from the mid-1960s through
Author Affiliations
1. Physical Activity and Public Health, School of Sport, Exercise and Health Sciences,
Loughborough University, UK
2. Community Health and Epidemiology, College of Medicine, University of
Saskatchewan, Saskatoon, SK
Correspondence: Nazeem Muhajarine, PhD, Community Health and Epidemiology,
College of Medicine, University of Saskatchewan, 107 Wiggins Rd., Saskatoon, SK
S7N 5E5, E-mail: [email protected]
Acknowledgements: The study was funded by the Heart and Stroke Foundation of
Canada, the Canadian Institutes of Health Research (Institute of Nutrition, Metabolism
and Diabetes) and the Rx&D Health Research Foundation. We thank the Saskatoon
school boards (Saskatoon Public and Greater Saskatoon Catholic Schools) for their
cooperation and the study participants for generously contributing their time. We
appreciate the work of research staff, including Tracy Ridalls. The Smart Cities, Healthy
Kids research team: Nazeem Muhajarine (Principal Investigator), Adam Baxter-Jones,
Scott Bell, Karen Chad, Charlie Clark, Dale Esliger, Paul Hanley, Bill Holden, Sara Kirk,
Cory Neudorf, Lauren Sherar and Lan Vu.
PHYSICAL ACTIVITY AND NEIGHBOURHOOD DESIGN
Table 1.
Accelerometry Data Collection and Analytical Procedures: Smart Cities, Healthy Kids Study, Saskatoon, Canada
Information
Accelerometer Model*
Piezo sensor orientation
Mode setup
Epoch
Deployment method
Location worn
Requested days of wear
Initialization
Wear instructions
Details
Actical (Philips Respironics, Bend, OR)
Omnidirectional
Counts only
15 seconds
Delivered and attached by researcher (on day 0)
Right hip at mid clavicular line (via adjustable neoprene waist belt)
7 d (40,320 epochs) not including day 0
Delayed until next day (i.e., day 1 at midnight 00:00 hrs)
Wear during waking hours, remove for sleeping†
Analytical
Non-wear appropriation
Valid day criteria
Valid file
Missing data
Cut-point reference(s)
Sedentary
MVPA
60 min (240 epochs) of consecutive 0s allowing for 2 min (8 epochs) of interruptions
10 h of wear
At least 1 valid file
No data modeling or imputation was performed
From Colley and Tremblay16 and Colley et al.17
<100 counts per minute
≥1500 counts per minute
* The Actical model has been validated in children.16
† We also asked that the accelerometer be removed for water-based activities where there was a chance of the device being fully submerged at a depth of
≥1 metre for ≥30 minutes (IEC Standard 60529 IPX7).
MVPA=moderate-to-vigorous physical activity.
to1998 and are characterized by their moderate density (10.64 people/
acre), and almost exclusively residential and highly car-oriented
configurations. Although auto-centric, the curving streets,
cul-de-sacs and intersections also contribute to lower throughtraffic. These three distinct Saskatoon neighbourhood designs,
which reflect the urban development approaches prevailing at the
time of their construction and are commonly found in other large
cities throughout Canada, can be used to assess whether particular
urban designs create built environments that are conducive to children being more active and/or less sedentary.
At the same time that great strides are being made in our understanding of the built environment, our ability to measure physical
activity is also evolving. Perhaps the most significant advance has
been the increased use of objective measurements of physical activity (e.g., with accelerometers) in the place of self-report measures,
which has greatly enhanced our ability to accurately capture short
bouts of lifestyle-embedded physical activity that are difficult to
recall.13 Advanced data-reduction software now enables more
detailed information on physical activity and time spent in sedentary behaviour to be derived from accelerometer data and, thus,
allows movement frequency, intensity, duration and temporality
to be investigated more comprehensively.14,15
Built environment policies are unfolding and evolving continually across the country, and are largely planned and funded by private, community, non-profit and/or public organizations both
within and outside the health sector. However, these “natural
experiments” often lack the embedded mechanisms needed for rigorous examination of how the multiple social, cultural, economic
and/or environmental factors affect the outcomes. These natural
experiments represent a rich but untapped data source that could
benefit from researcher involvement and expertise. Therefore, the
purpose of this study was to provide a detailed accelerometry-based
profile of physical activity and sedentary time in children living in
the three differing neighbourhood design types: 1) core, 2) fracturedgrid and 3) curvilinear pattern. The overall objective was to
determine whether, and to what extent, a relation exists between
neighbourhood design, as classified here, and children’s physical
activity and sedentary time in Saskatoon. While we recognize that
the three neighbourhood design types represent a detailed and rich
mix of various specific built environment characteristics, not all of
Figure 1.
Saskatoon neighbourhood planning eras, Smart
Cities, Healthy Kids study
which show a linear pattern with neighbourhood type, we hypothesize that, in the main, children’s physical activity level will fall
and their sedentary behaviour will increase as we move from the
core to the fractured-grid to the curvilinear neighbourhoods. A secondary hypothesis was that the pattern and timing of physical
activity (i.e., weekdays vs. weekend days and time of day) of the
children would differ depending on the design of the neighbourhood in which they live.
METHODS
The study uses data from Smart Cities, Healthy Kids, an ongoing
mixed-method, multi-phase population health intervention study.
Smart Cities recruited 1,610 children from grades 5 to 8 (10-14
years) from 40 elementary schools, and these children represented
all the 60 residential neighbourhoods of Saskatoon as of 2010.
Saskatoon is the largest city in the prairie province of Saskatchewan
and has a metropolitan area population of 233,893 according to
the 2006 census. The participants selected included a subsample of
455 children (52% female) from the larger study, who were requested to wear accelerometers for 7 days.
Table 2.
Descriptive Characteristics of the Study Sample by Neighbourhood Design Type
Variable
Sampled schools, n
Family income, % in group 1-2*
Weekday compliance (n=432)
Weekend compliance (n=349)
Mean (SD) age, decimal years
Sex, % girls
Mean (SD) body mass index
Min; Max
OW & OB, %
Mean (SD) weekday wear, minutes
Min; Max
Mean (SD) weekend day wear, minutes
Min; Max
Mean (SD) counts per minute
Min; Max
ICC weekday counts
ICC weekend day counts
Mean (SD) MVPA, min/day
Min; Max
Mean (SD) sedentary time, min/day
Min; Max
Core
6
70.2
121
97
11.6 (1.2)
54
20.1 (4.7)
14.1; 42.4
11.6
797 (59)
612; 961
758 (87)
609; 1002
460 (202)
94;1330
0.656
0.540
77.5 (37.5)
8.1; 232.5
528 (73)
317; 683
Fractured-Grid
10
71.5
137
107
11.6 (1.2)
51
20.4 (4.3)
13.4; 34.3
14.0
801 (63)
622; 947
754 (90)
601;1054
403 (164)†
118;1225
0.766
0.605
67.7 (31.3)†
13.3; 234.6
543 (72)
344; 760
Curvilinear-Pattern
14
81.9
174
145
11.8 (1.2)
52
20.0 (3.8)
13.2; 34.0
16.8
808 (55)
667; 933
761 (90)
605;1124
441 (169)
127;1086
0.760
0.509
73.3 (29.2)
16.6; 182.0
537 (63)
331; 798
* Family income groups: 1 – Wealthy, 2 – Average, 3 – Difficult, 4 – Poor.
† Indicates significant group difference (p<0.05) compared with both core and curvilinear-pattern neighbourhood design types.
OW & OB=overweight and obesity combined, calculated using age-specific cut-offs;21 ICC=intra-class correlation; MVPA=moderate-to-vigorous physical activity.
Analytically, the children were grouped according to the design
of the neighbourhood in which they lived (Figure 1): Era 1 – core
(N=127), Era 2 – fractured-grid (N=146) and Era 3 – curvilinearpattern (N=182). Each child gave written assent, and parental written
informed consent was also obtained. All procedures were approved
by the institutional research ethics board.
A portable stadiometer and weigh scale were used to measure
height and weight. Body mass index (BMI) was calculated from
height and weight (weight [kg]/height [m]2). Physical activity and
sedentary time were objectively measured for seven consecutive
days by means of an Actical accelerometer (Mini Mitter Co., Inc.,
Bend, OR, US). All pertinent data collection and analytical procedures related to the accelerometry portion of the study are
described in Table 1. The raw data were analyzed using custom software KineSoft, version 3.3.63 (KineSoft, SK) to produce a series of
standardized outcome variables following procedures similar to
those described by Esliger et al.14,15 The main variables of interest
were average counts per minute, minutes of moderate-to-vigorous
physical activity (MVPA) and minutes spent sedentary. In an effort
to understand the temporality of the physical activity and time
spent sedentary, these variables were analyzed over the following
time periods: weekday, weekend day and hourly.
Studies have shown a positive association between socio-economic
status (SES) and MVPA18,19 and an inverse association between
SES and time spent in sedentary behaviour among children.20 It is
therefore important to control for differences in SES at the individual level. Thus in all multivariate analyses, family income was
controlled for in addition to age, sex and accelerometer wear time.
Children were asked “Would you describe your family money situation as…” indicating one response in a 4-point Likert scale
(1 – Wealthy, 2 – Average, 3 – Difficult, 4 – Poor). The responses to this
question corresponded with the answers to the questions regarding
the mother’s/father’s highest level of schooling, as was expected.
One-way ANOVA models were used to test for group differences
in chronological age and accelerometer wear time. Subsequent
analyses used multivariate analysis of covariance (MANCOVA)
models with chronological age, income and wear time as the
covariates to determine group differences in physical activity and
sedentary variables. Owing to the skewed distribution of the MVPA
variables, these data were log transformed. All statistical tests were
performed on the transformed data; however, the non-transformed
means and standard deviations are presented. The influence of time
of day on physical activity and sedentary time was described visually in 24 h x 7 d area plots for each group. Where appropriate,
models used Fisher’s least significant difference for post-hoc comparisons, and alpha was set at p<0.05. All analyses were performed
using SPSS for Windows version 20.0 (SPSS Inc., Chicago, IL).
RESULTS
The characteristics of the study sample (N=455) are displayed by
neighbourhood design in Table 2. No significant group differences
were found for age, sex, BMI, prevalence of overweight and obesity
combined, weekday or weekend day wear time. The sample size
retained after removing those found to be noncompliant (i.e., not
wearing the accelerometer for at least 10 hours/day on 1 or more
days), as well as those with missing covariates, was 432 for the weekday analyses and 349 for the weekend day analyses (95% and 77%,
respectively). In addition to the excellent compliance rates, the three
groups greatly exceeded the minimum 10 hours/day wear-time
requirement with an overall average of 13 hours, 11 minutes per day.
When age, sex and family income had been controlled for,
ANCOVA models revealed that children living in fractured-grid
neighbourhoods had fewer average accelerometer counts per
minute on weekdays than the core and curvilinear-pattern neighbourhoods (83 and 55 minutes less, respectively) (Figure 2, top).
Further analyses using MANCOVA models revealed that, after age,
sex, family income and accelerometer wear time had been controlled for, children in the fractured-grid neighbourhoods accumulated significantly less MVPA (15 and 9 minutes) per weekday
and spent a significantly greater time in sedentary behaviour
(23 and 17 minutes) than those in core and curvilinear-pattern
neighbourhoods, respectively (Figure 2, top). No significant group
differences were evident for counts per minute, MVPA or sedentary
time on the weekend days (Figure 2, bottom).
Figure 2.
Average weekday physical activity and time spent
sedentary (top) and weekend day physical activity
and time spent sedentary (bottom) by
neighbourhood eras
Figure 3.
Sedentary
Light
MVPA
60
100
700
Group differences (Top=Core, Middle=Fracturedgrid, Bottom=Curvilinear-pattern) in intensity-specific
physical activity profiles (7 d × 24 h)
90
50
*
80
*
70
*
60
400
50
300
40
Minutes Per Weekday
Counts Per Minute
Minutes Per Weekday
500
30
200
Minutes Spen
nt at a Given Intensity Per Hour
600
40
30
20
10
Core
20
100
10
M
0
0
Sedentary
T
18
0
6
12
W
18
0
6
12
T
18
0
6
12
F
18
0
6
12
18
S
0
6
12
18
S
0
6
12
0
6
12
0
6
12
18
Hour of the Day
CP
FG
Core
CP
FG
Core
CP
FG
Counts Per Minute
Sedentary
Light
MVPA
60
MVPA
700
100
600
80
500
60
400
50
300
40
30
200
Minutes Per W
Weekend Day
70
Minutes Spen
50
90
Counts P
Per Minute
Minutes Per Weekend Da
ay
12
0
Core
0
6
40
30
20
10
Fracture Grid
M
0
20
0
6
12
T
18
0
6
12
W
18
0
6
12
18
T
0
6
12
F
18
0
6
12
S
18
0
6
12
18
S
18
Hour of the Day
100
10
Sedentary
Light
MVPA
60
CP
C
FG
F
Co
ore
CP
C
FG
F
Sedentary
MVPA
Means adjusted for sex, age, family income and wear-time covariates;
FG=fractured-grid; CP=curvilinear-pattern; MVPA=moderate-to-vigorous
physical activity; MVPA is scaled according to the secondary y axis. Analyses
were performed on log-transformed physical activity variables; however, nontransformed data are presented. Error bars represent standard error.
*Fractured-grid group significantly different from both the core and
curvilinear-pattern groups (p<0.05).
In terms of temporality, Figure 3 shows the hourly physical activity behaviours by intensity and neighbourhood design. Focusing
on school days (i.e., Monday to Friday from 0800 to 1500 h), it
appears that the children living in different neighbourhoods had
similar patterns of physical activity and sedentary time (i.e., they
have the same physical activity intensity “hotspots”). Overall, on
weekends, children spent 70% of their days being sedentary
(528 minutes), whereas on weekdays sedentary time accounted for
67% of their day (540 minutes). Overall, on weekends, children
spent 7% of their day (52 minutes) accumulating MVPA compared
with 10% of the day (80 minutes) on weekdays. In terms of temporality, all three groups had a quadra-modal distribution (i.e., four
pronounced peaks) of MVPA: during the morning commute (hour
8), morning recess (hour 10), lunch (hour 12) and afternoon commute (hour 15) (Figure 4, top). Comparison of the three peak sedentary hours of the school day (i.e., hour 9, hour 11 and hour 13)
50
Minutes Spen
Counts Per Minute
Co
ore
CP
C
FG
F
0
Co
ore
0
40
30
20
Curvilinear-Pattern
10
M
0
0
6
12
T
18
0
6
12
W
18
0
6
12
18
T
0
6
12
F
18
0
6
12
18
S
0
6
12
18
S
18
Hour of the Day
Note: White areas of the figure (i.e., from approximately midnight to
approximately 6 am) represent non-wear (sleep) time.
showed that children from the core, fractured-grid and curvilinearpattern neighbourhoods were sedentary, on average, 44, 46 and 45
of every 60 minutes, respectively, during these three 1-hour blocks
(Figure 4, bottom). Compared with the in-school time, the children’s out-of-school sedentary time (i.e., Monday to Friday from
1500 to 2100 h) increased by 1 minute per hour, and their time
spent in MVPA decreased by 1 minute per hour (Figure 4).
Although children in the fractured-grid neighbourhoods were
significantly less active than children living in the core and the
Figure 4.
Average minutes spent in MVPA (top) and
sedentary behaviour (bottom) hour by hour on a
weekday, by neighbourhood era
10
Average
e Minutes of MVPA Per Hour
8
6
4
2
Era 1 - Core
Era 2 - Fractured Grid
Era 3 - Curvilinear-Pattern
0
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
21
22
23
Weekday Hour
Average Minutes Spent Sedentary Per Hour
50
45
40
35
Era 1 - Core
Era 2 - Fractured Grid
Era 3 - Curvilinear-Pattern
30
7
8
9
10
11
12
13
14
15
16
17
18
19
20
Weekday Hour
curvilinear neighbourhoods, visually the patterns of weekday physical activity are similar across the groups. The magnitude of this
difference, per hour, may appear to be small (e.g., ~1 minute less
MVPA per hour), but when accumulated over the entire day it
results in a meaningful difference (Figure 4, top).
DISCUSSION
The purpose of this study was to determine whether, and to what
extent, a relation exists between neighbourhood design (i.e., core,
fractured-grid or curvilinear-pattern) and children’s physical activity and time spent being sedentary, in Saskatoon. This observational study supports the hypothesis that children 10-14 years of
age living in a fractured-grid neighbourhood are likely to be less
active (fewer minutes in MVPA) and more sedentary than children
in core or curvilinear neighbourhoods. The physical activity levels
of the overall sample (i.e., 73 minutes of MVPA per day) is higher
than in other, similarly analyzed population samples (e.g., Canadian Health Measures Survey (CHMS); 58 minutes MVPA per day in
children 6 to 14 years17). The higher MVPA in the present sample
may be due, among other things, to the fact that the data were collected during the fair weather months of April-June and did not
include the typically less active winter months, as the CHMS did.
When minutes spent in sedentary, light and MVPA are presented for the full week (Figure 3), it is apparent that despite the lower
MVPA and higher sedentary time in the fractured-grid neighbourS26 REVUE CANADIENNE DE SANTÉ PUBLIQUE • VOL. 103, SUPPLÉMENT 3
hoods, the patterns of behaviours are markedly similar among all
three neighbourhood designs. For example, children from each
neighbourhood group were more active on weekdays compared
with weekend days, a finding that is supported in the literature.22,23
Furthermore, on average, children from each neighbourhood type
showed the same weekday activity “hotspots” before and after
school, and during recess and lunch (Figure 4). These observations
suggests that although the neighbourhood design has an appreciable impact on children’s overall time spent in MVPA, the general
physical activity pattern, hourly, on a typical weekday is not dramatically different across neighbourhood types. This shows the
important contribution that schools make to patterning children’s
physical activity, whether it is related to school travel or activity
levels during time in school.
Our research showed that children in the fractured-grid pattern
neighbourhoods accumulated 15 and 9 minutes less MVPA per day
than children in the core and curvilinear-pattern neighbourhoods,
respectively. Emerging research has shown that modest differences
in accelerometer-assessed MVPA, similar to the differences shown
in this study, are associated with health benefits in youth. For
example, data from a larger sample (N=5,500) of 12-year-olds
showed that an increase of 15 minutes of daily MVPA was associated with lower odds of obesity – over 50% in boys and nearly 40%
in girls.24
Children living in the core neighbourhoods accumulated the
greatest minutes of MVPA (although not significantly greater than
the curvilinear-pattern neighbourhood). Core neighbourhoods are
less car-oriented and have greater land-use mix (combining commercial, residential, institutional land uses). A recent systematic
review6 revealed land-use mix to be a strong correlate of physical
activity among children and adolescents. In addition, the gridpatterned road networks found in the core neighbourhoods with
multiple intersections and interconnected streets increased the
number of access and exit routes into a neighbourhood. This provides greater route choice and accessibility to destinations. Past
research has shown that children from neighborhoods with greater
street connectivity are more active outside their back/front yard
when compared with children who live in curvilinear (cul-de-sac)
neighbourhood designs.25 However, it should be noted that
increased connectivity of streets may also be coupled with increased
traffic on local roads (because of vehicle use of short cuts).26 Traffic
speed/volume in neighbourhood streets is an important correlate of
reduced physical activity in children (due to safety issues),6 thus
street connectivity could also be a barrier to physical activity in the
core neighbourhoods. This highlights the complexity of the association between neighbourhood design – as illustrated by street
connectivity here, for example – and children’s physical activity.
Children in the fractured-grid neighbourhood designs had significantly less MVPA and significantly greater sedentary time than
children from the core and curvilinear-pattern neighbourhoods.
Fractured-grid neighbourhoods are more reliant on car transport
(i.e., auto-centric) and have lower density of destinations than the
core neighbourhoods, which may contribute to the lower activity
of the children residing in those neighbourhoods.27 However, fracturedgrid neighbourhoods include more residential land use than found
in the core neighbourhoods, and residential density has been
shown to be positively associated with self-reported physical activity in children.6 The fractured-grid neighbourhoods in Saskatoon
are usually bordered by busy, high-traffic arterial roads with little
pedestrian access; this too could reduce the walkability of these
neighbourhoods and likely the roaming distance of children.
Increasingly, Canadian cities, including Saskatoon, are adopting
a newer neighbourhood design pattern called the Fused-Grid Model,
which incorporates features from the core and the curvilinearpattern neighbourhoods.28 The Fused-Grid Model, similar to the
core neighbourhoods, provides commercial destinations that can
be easily accessed by foot but incorporates the traffic calming measures of curving streets, cul-de-sacs and intersections seen in the
curvilinear-pattern neighbourhoods. The overall aim of the fusedgrid design is to promote active transportation, increase the opportunities for social interaction within the neighbourhood and
improve traffic safety.28
A strength of this study was the use of objective measurements
of physical activity.13 Another strength was the detailed analysis of
the accelerometry data in terms of intensity and temporal patterns.
However, there are limitations to accelerometry, most notably, its
inability to assess lifting, carrying, cycling and water-based activities, and the general lack of contextual information relating to
activity mode and/or location/domain.29 For example, the fact that
waist-mounted accelerometers do not measure cycling could have
limited the true quantification of MVPA in neighbourhoods that
are more “bike friendly”.
Although data across all neighbourhoods were collected during
one season (April 28-June 11), data on extreme weather were not
considered in these analyses. However, the likelihood that acute
weather would have systematically affected the results is minimal.
Another limitation was that selection bias could not be ruled out
because of the non-random nature of the sample. The selfidentifying process of school selection could have resulted in
schools participating that were more supportive of physical activity. Last, the cross-sectional design of the study and the inability to
control for self-selection or sorting into a residential area prevents
comment on causation. A broad classification of neighbourhood
design (i.e., core, fractured-grid, curvilinear-pattern) was chosen
because these are the district neighbourhood designs present in
Saskatoon and resonate with policy-makers, developers and city
planners. Also, these design types typically reflect the era or the
vintage of neighbourhoods and are commonly found in other cities
across Canada.
Our research suggests that the neighbourhood design is associated with differences in MVPA and time spent being sedentary.
However, we do not have information on the location of the activity, and therefore it is not known whether the activity took place in
the neighbourhood in which the child resided or away from it. For
example, a child residing in the core neighbourhoods may be driven to a fractured-grid neighbourhood for organized sports (e.g., soccer), as indicated by some participants in a qualitative substudy
derived from this same sample (data not shown). Global positioning systems (GPS) provide time-stamped location information and
would tell us in which neighbourhood the MVPA occurred.30
Research has started to emerge that examines the associations
between aspects of the built environment and physical activity in
youth using both accelerometry and GPS;2-5 however, the
accelerometer data are limited to day-level aggregate variables (such
as daily minutes in MVPA) without any investigation as to where
and how the activity and sedentary behaviour occur.14,15 Future
research is warranted that employs a detailed profiling of MVPA
and sedentary behaviour by accelerometry, and research that couples it with GPS data collection.
In conclusion, the data presented in this paper show that in
Saskatoon there is a relation between neighbourhood design and
children’s physical activity and sedentary behaviours. Further work
will need to be done to tease out which aspects of the built environments specifically have the greatest impact on these behaviours.
This information, offered in the context of ongoing development
of neighbourhoods, as we see in Saskatoon, is critical to an
evidence-informed approach to modern urban development and
planning.
REFERENCES
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
Cooper AR, Page AS, Wheeler BW, Hillsdon M, Griew P, Jago R. Patterns of GPS
measured time outdoors after school and objective physical activity in English children: The PEACH project. Int J Behav Nutr Phys Act 2010;7:31.
Sundquist K, Eriksson U, Kawakami N, Skog L, Ohlsson H, Arvidsson D.
Neighborhood walkability, physical activity, and walking behavior: The
Swedish Neighborhood and Physical Activity (SNAP) study. Soc Sci Med
2011;72(8):1266-73.
Maddison R, Hoorn SV, Jiang Y, Mhurchu CN, Exeter D, Dorey E, et al. The
environment and physical activity: The influence of psychosocial, perceived
and built environmental factors. Int J Behav Nutr Phys Act 2009;6:19.
Rodriguez DA, Cho GH, Evenson KR, Conway TL, Cohen D, Ghosh-Dastidar B,
et al. Out and about: Association of the built environment with physical activity behaviors of adolescent females. Health Place 2012;18(1):55-62.
De Meester F, Van Dyck D, De Bourdeaudhuij I, Deforche B, Sallis JF, Cardon G.
Active living neighborhoods: Is neighborhood walkability a key element for
Belgian adolescents? BMC Public Health 2012;12:7.
Ding D, Sallis JF, Kerr J, Lee S, Rosenberg DE. Neighborhood environment and
physical activity among youth: A review. Am J Prev Med 2011;41(4):442-55.
Ashe M, Graff S, Spector C. Changing places: Policies to make a healthy choice
the easy choice. Public Health 2011;125(12):889-95.
Srinivasan S, O’Fallon LR, Dearry A. Creating healthy communities, healthy
homes, healthy people: Initiating a research agenda on the built environment and public health. Am J Public Health 2003;93(9):1446-50.
Lavizzo-Mourey R, McGinnis JM. Making the case for active living communities. Am J Public Health 2003;93(9):1386-88.
Dannenberg AL, Jackson RJ, Frumkin H, Schieber RA, Pratt M, Kochtitzky C,
et al. The impact of community design and land-use choices on public health:
A scientific research agenda. Am J Public Health 2003;93(9):1500-8.
Durand CP, Andalib M, Dunton GF, Wolch J, Pentz MA. A systematic review
of built environment factors related to physical activity and obesity risk:
Implications for smart growth urban planning. Obes Rev 2011;12(5):e173e182.
City of Saskatoon Urban Design – Land Branch. Public Spaces, Activity and
Urban Form: Strategic Framework – City Centre Plan Phase 1. 2011.
Shephard RJ. Limits to the measurement of habitual physical activity by questionnaires. Br J Sports Med 2003;37(3):197-206.
Esliger DW, Tremblay MS, Copeland JL, Barnes JD, Huntington GE,
Bassett DR, Jr. Physical activity profile of Old Order Amish, Mennonite, and
contemporary children. Med Sci Sports Exerc 2010;42(2):296-303.
Esliger DW, Tremblay MS. Physical activity and inactivity profiling: The next
generation. Can J Public Health 2007;98(Suppl 2):S195-S207.
Colley RC, Tremblay MS. Moderate and vigorous physical activity intensity
cut-points for the Actical accelerometer. J Sports Sci 2011;29(8):783-89.
Colley RC, Garriguet D, Janssen I, Craig CL, Clarke J, Tremblay MS. Physical
to 2009 Canadian Health Measures Survey. Health Rep 2011;22(1):15-23.
Gordon-Larsen P, McMurray RG, Popkin BM. Determinants of adolescent
physical activity and inactivity patterns. Pediatrics 2000;105(6):E83.
Butcher K, Sallis JF, Mayer JA, Woodruff S. Correlates of physical activity
guideline compliance for adolescents in 100 U.S. Cities. J Adolesc Health
2008;42(4):360-68.
Jago R, Page A, Froberg K, Sardinha LB, Klasson-Heggebo L, Andersen LB.
Screen-viewing and the home TV environment: The European Youth Heart
Study. Prev Med 2008;47(5):525-29.
WHO Multicentre Growth Reference Study Group. WHO Child Growth Standards based on length/height, weight and age. Acta Paediatr Suppl
2006;450:76-85.
Nader PR, Bradley RH, Houts RM, McRitchie SL, O’Brien M. Moderate-tovigorous physical activity from ages 9 to 15 years. JAMA 2008;300(3):295-305.
Treuth MS, Catellier DJ, Schmitz KH, Pate RR, Elder JP, McMurray RG, et al.
Weekend and weekday patterns of physical activity in overweight and
normal-weight adolescent girls. Obesity (Silver Spring) 2007;15(7):1782-88.
24. Ness AR, Leary SD, Mattocks C, Blair SN, Reilly JJ, Wells J, et al. Objectively
measured physical activity and fat mass in a large cohort of children. PLoS Med
2007;4(3):e97.
25. Holt NL, Spence JC, Sehn ZL, Cutumisu N. Neighborhood and developmental differences in children’s perceptions of opportunities for play and physical activity. Health Place 2008;14(1):2-14.
26. Handy S, Paterson RG, Butler K. Planning for Street Connectivity: Getting
from Here to There. Chicago, IL: American Planning Association, 2003.
27. Mackett R, Brown B, Gong Y, Paskins J. Children’s independent movement in
the local environment. Built Environment 2007;33:454-68.
28. Grammenos F. The fused grid: A contemporary urban pattern. Available at:
http://www fusedgrid ca/contactus php 2008 (Accessed February 2, 2012).
29. Montoye HJ, Kemper HCG, Saris WHM, Washburn RA. Measuring Physical
Activity and Energy Expenditure. Champaign, IL: Human Kinetics, 1996.
30. Jones AP, Coombes EG, Griffin SJ, van Sluijs EM. Environmental supportiveness for physical activity in English schoolchildren: A study using global positioning systems. Int J Behav Nutr Phys Act 2009;6:42.
RÉSUMÉ
Objectifs : Déterminer s’il existe une relation, et si oui de quelle
ampleur, entre, d’une part, la conception du quartier et, d’autre part,
l’activité physique et les comportements sédentaires des enfants à
Saskatoon.
Méthode : Trois types de quartiers ont été évalués : 1) les quartiers du
centre-ville datant d’avant 1930, à l’agencement quadrillé, 2) les
quartiers scindés à agencement quadrillé datant des années 1930 au
milieu des années 1960 et 3) les quartiers à agencement curviligne
datant du milieu des années 1960 à 1998. Nous avons mesuré
objectivement par accélérométrie, pendant 7 jours, l’activité physique et
les comportements sédentaires d’enfants de 10 à 14 ans (N=455; âge
moyen 11,7 ans), regroupés selon leur quartier domiciliaire. Des modèles
ANCOVA et MANCOVA (analyse multivariée de la covariance) ont servi à
l’évaluation des différences entre les groupes (p<0,05).
Résultats : Nous avons constaté des différences entre les groupes les
jours de semaine, mais non les samedis et dimanches. Après ajustement
selon l’âge, le sexe et le revenu familial, les enfants habitant des quartiers
scindés à agencement quadrillé enregistraient en moyenne 83 et 55
points d’accéléromètre de moins par minute les jours de semaine que les
enfants des quartiers du centre-ville et des quartiers à agencement
curviligne, respectivement. Une analyse plus poussée a montré que les
enfants des quartiers scindés à agencement quadrillé accumulaient 15 et
9 minutes de moins d’activité physique modérée à vigoureuse par jour et
consacraient plus de temps à des comportements sédentaires (23 et 17
minutes) que ceux des quartiers du centre-ville et des quartiers à
agencement curviligne, respectivement.
Conclusion : Ces données montrent qu’à Saskatoon, il y a une relation
entre la conception des quartiers et l’activité physique et les
comportements sédentaires des enfants. Il faudrait pousser la recherche
pour « démêler » quelles caractéristiques des milieux bâtis ont le plus
d’impact sur ces importants comportements liés au mode de vie. Cette
information, présentée dans le contexte du développement en cours des
quartiers, comme on le voit à Saskatoon, est essentielle à une démarche
de planification et de développement urbain fondée sur des preuves.
Mots clés : urbain; milieu bâti; accéléromètre; mode de vie; urbanisme
QUALITATIVE RESEARCH
Walkable for Whom? Examining the Role of the Built Environment
on the Neighbourhood-based Physical Activity of Children
Kristjana Loptson, MA,1 Nazeem Muhajarine, PhD,1,2 Tracy Ridalls, MA1 and the Smart Cities, Healthy Kids
Research Team*
ABSTRACT
Objectives: To date, only a few studies have attempted to study the processes by which community design and the built and social environments
affect individual physical activity, especially in children. Qualitative enquiry is useful for exploring perceptions and decision-making, and to understand
the processes involved in how people interact with their environments. This study used qualitative methods to gain insight into the pathways linking the
neighbourhood environment with children’s activity patterns.
Methods: Data were collected in semi-structured interviews with 24 child-parent dyads (children aged 10-14 years). Families lived in neighbourhoods
ranging from lowest to highest median income and representing the three main design types found in Saskatoon – urban, semi-suburban and
suburban.
Results: Parents and children underscored the importance of safe environments for children’s physical activity: streets or paths they can cycle on
without feeling threatened, parks and green spaces free of criminal activity, and neighbourhoods where people know each other and children have
friends to play with. Although grid-pattern urban neighbourhoods with a high density of destinations may in principle promote active transportation,
the higher levels of crime and traffic danger that tend to exist in these areas may hinder physical activity in children.
Conclusion: Understanding what facilitates activity in children is a complex endeavour. It requires understanding the barriers to physical activity
present at the neighbourhood level as well as social and perceptual factors that act in interdependent ways to either promote or hinder children’s
physical activity.
Key words: Neighbourhood built environment; children; qualitative method; safety; physical activity
W
alkability – the extent to which an area is supportive of
walking – is a concept that emerged from the transportation literature and has been widely adopted in
health research examining the impact of the built environment on
physical activity and health outcomes.1 Factors that make neighbourhoods more walkable include pedestrian amenities such as
sidewalks, crosswalks, curb cuts and traffic lights; street connectivity; mixed-land use; and the presence of a variety of destinations
within walking distance, features typically found in urban more
than suburban neighbourhoods.2-4 From a public health perspective, creating more walkable neighbourhoods might be expected to
lead to a healthier environment by encouraging reduced car usage
and therefore lower car emissions and air pollution, and also by
increasing opportunities for active transportation (physically active
modes of transportation, such as walking, biking, rollerblading,
skateboarding), which could increase overall levels of physical
activity and decrease obesity.5-7
Although a significant amount of research has shown that adults
living in urban neighbourhoods walk more and have a lower bodymass index (BMI) than their suburban counterparts, other studies
have found that this association is not consistent in all urban
neighbourhoods or with all demographic groups.4,8,9 Very little
research has examined the impact of neighbourhood design on
activity levels in children and youth, and the few studies that have
looked specifically at youth activity have also produced mixed findings.10-13 A study of Belgian adolescents found that they were more
likely to walk and bike in less walkable neighbourhoods than more
walkable neighbourhoods.14 Other studies have found that while
boys are more active in neighbourhoods that are close to commercial areas and have connected streets, girls are more active in neighbourhoods with unconnected, curvilinear, low-traffic streets.15,16
No consistent association has been established between children’s BMI and neighbourhood design, but some research suggests
that certain neighbourhood characteristics may be influential. For
example, neighbourhood safety and access to parks, playgrounds,
recreation centres and sidewalks were significantly associated with
lower BMI in girls aged 10-11 years in a US study based on a survey
conducted by the National Centre for Health Statistics.17 Higher
rates of overweight and obesity were found in both boys and girls
Author Affiliations
1. Saskatchewan Population Health and Evaluation Research Unit, SK
2. Department of Community Health and Epidemiology, College of Medicine,
University of Saskatchewan, Saskatoon, SK
* Smart Cities, Healthy Kids research team: Nazeem Muhajarine (Principal
Investigator), Karen Chad, Cory Neudorf, Adam Baxter-Jones, Bill Holden,
Scott Bell, Charlie Clark, Lauren Sherar, Dale Esliger, Sara Kirk, Paul Hanley and Lan Vu.
Correspondence: Nazeem Muhajarine, PhD, Community Health and Epidemiology,
College of Medicine, University of Saskatchewan, 107 Wiggins Rd., Saskatoon, SK
S7N 5E5, E-mail: [email protected]
Canada, the Canadian Institutes of Health Research (Institute of Nutrition, Metabolism
and Diabetes) and the Rx&D Health Research Foundation. We thank the Saskatoon
school boards (Saskatoon Public and Greater Saskatoon Catholic Schools) for their
cooperation and the study participants for generously contributing their time. We
appreciate the work of the research staff.
NEIGHBOURHOODS AND CHILDREN’S PHYSICAL ACTIVITY
living in neighbourhoods perceived to be unsafe or where garbage
and other signs of neighbourhood disorder were evident.17
Despite the wide adoption of walkability measurements in the
study of neighbourhoods, numerous questions remain to be
answered about this concept, especially its relevance to children.
The amount of neighbourhood-based physical activity is likely to
be influenced not only by the walkability of an area as it is typically
measured but also by the neighbourhood’s socio-economic status
(SES) and the characteristics of the individual.18,19 Studies from both
Canada and the United States have found that BMI tends to be
higher in low-income neighbourhoods, despite their typically high
“walkability”, suggesting that social factors may modify the relation between the built environment and behaviour.4,20,21
Given the complications in capturing complex neighbourhood
influences on individual behaviour, understandably few attempts
to date have been made to investigate the processes by which the
physical and social environments affect individual adaptations,
especially in children. The study reported here used qualitative
methods to gain insight into the processes linking the neighbourhood environment with children’s activity patterns.
METHODS
This study comes from the third and final phase of the Smart Cities,
Healthy Kids project in Saskatoon, SK, a city of 240,000 people. In
2010, 24 children in Grades 5-8, representing a range of residential
neighbourhood types in Saskatoon, and the mothers of the children were interviewed to find out what influence they felt their
neighbourhood had on the children’s activity levels.
The 455 families whose children had worn accelerometers in an
earlier phase of our study were invited to take part in this qualitative component. We used purposive sampling to achieve representativeness of neighbourhood types, and of those families that
volunteered we selected 24 living in 18 neighbourhoods, with the
goal of having the full range of neighbourhood designs represented. Before the interviews, each child was loaned a camera to take
pictures of aspects of the built environment that they felt either
encouraged or discouraged their physical activity. Semi-structured,
in-person interviews were conducted separately with the child and
with one of the child’s parents. In all cases, this ended up being the
mother (or in one instance, the step-mother); in two interviews,
fathers were also present. The photos the child had taken were used
as a starting point for talking with him or her about the impact of
the neighbourhood on activity (but not in the parental interview).
The interviewers, two of the authors of this paper, have social science backgrounds and extensive experience in qualitative research.
The interviews covered a range of questions, including perceptions
of safety, general feelings about the neighbourhood, barriers to
physical activity, time management and exercise habits.
Each interview was audio-recorded and transcribed. Using
NVIVO 9 (QSR Int.), the two interviewers created independent coding lists and then collaborated to create a master coding list based
on themes they had identified in the interviews.22,23 A long list of
themes emerged that related to a range of topics, including perceptions about the neighbourhood, school and work environments, screen time, family values and rules, financial
considerations, transportation habits, social and recreation preferences, concerns about safety, the role of gender, the presence of siblings and pets, and technology. Although many of the themes
appeared in both interview types, the interviews were coded according to whether they were child or parent interviews. The child and
parent interviews within each dyad were then cross-compared to
examine the points of similarity and divergence between child and
parent perceptions of neighbourhood features and barriers or aids
to physical activity.
This study included a diverse group of neighbourhoods, ranging
from lowest to highest median income and representing the three
main neighbourhood design types found in Saskatoon. Neighbourhood boundaries are designated by the City, and neighbourhood
SES was determined by neighbourhood demographic information
from the 2006 census, made available by the City of Saskatoon.24
The median household income in the city, based on the 2006 census, was $66,507. For the purposes of the study, low-income neighbourhoods were those with median household incomes below
$50,000, and high-income neighbourhoods were those with median household incomes above $85,000. Urban neighbourhoods, built
prior to the 1930s, have a traditional grid design, consisting of
straight, intersecting streets and back alleys; they typically have
higher population density and are of mixed use. Semi-suburban
neighbourhoods surround the urban core and were built between
1931 and 1966; they have a mix of grid-based and curvilinear streets,
are of lower density, predominantly residential, and are increasingly car-oriented as they are located further away from the urban centre. Suburban neighbourhoods, built after 1966, are on the periphery
of the city, follow curvilinear street patterns, are low-density, almost
exclusively residential and highly car-oriented. Ten of the participating families resided in urban neighbourhoods, eight in semisuburban neighbourhoods and six in suburban neighbourhoods.
Although the trend is not consistent across all neighbourhoods, generally speaking there is the least amount of park space in Saskatoon’s
low-income urban neighbourhoods and the greatest in the highincome suburban neighbourhoods. Suburban parks tend to be located in low-traffic, low-crime areas, whereas urban park space is closer
to busy traffic intersections and areas often perceived to be unsafe,
either because of the presence of strangers or a total absence of people (i.e., no surveillance). Semi-suburban parks tend to resemble suburban parks in that they are often relatively well equipped and
maintained spaces and are perceived to be safer from both traffic
and crime than most urban parks.
RESULTS
Participants talked about the factors influencing children’s use of
active transportation as well as other types of physical activity, both
within and outside their neighbourhood. While they sometimes
mentioned aspects of the built environment, social factors were
more likely to be cited, and the influences on children’s activity
were reported to be different from the factors that affect adults.
Active transportation and schools
A key reason that neighbourhood walkability is considered important is that it facilitates active transportation. For children, the most
frequent opportunity to use active transportation is travelling to
and from school. In this study, children’s use of active transportation to school was related to the type of neighbourhood in which
they lived, mostly because of the location of schools.
Of the children living in urban neighbourhoods, only one
walked to school consistently, primarily because most of these chil-
dren attended schools outside their neighbourhoods. Children living in suburban or semi-suburban neighbourhoods were much
more likely to attend local schools and walked or biked to school
some or all of the time. This was seen positively by some parents,
for interpersonal as well as health reasons:
[My daughter] walks back and forth [to school] and she’s coming home
at lunch, too…At the end of the day if time permits and schedules are
such that I can walk and go and meet her, I still do, and though she’s
in Grade Six... it’s a good end of the day chat time.
–Mother, semi-suburban middle SES neighbourhood
Even when the distance between home and school makes walking feasible, parents’ attitudes play a role in determining how often
children actually do walk.
[In winter] I just say, “Put on more clothes.” She still has to walk to
school. I’m not a parent that drives to school all the time. She’s a big
girl and I’m a mean mother.
–Mother, suburban high SES neighbourhood
Numerous factors may be taken into account when deciding
whether a child will walk to school or be driven on a given day, as
this mother explains:
[Whether or not my children walk or bike to school] does depend on the
season. It also depends on whether it’s a band day or not, because biking with a big saxophone doesn’t really work well. We also consider
their afterschool activity; if they have two-and-a-half hours of sport
in the evening, I don’t push them to bike.
–Mother, urban middle SES neighbourhood
While children’s use of active transportation is most likely to be
limited to their neighbourhood, other types of physical activity can
and do occur outside the neighbourhood they reside in. However,
this usually requires transportation by parents. A number of participants reported leaving their neighbourhood by car to walk or
bike in a more desirable area.
It’s difficult biking safely with children around our neighbourhood. I
have been almost pushed off my bike by people...I would not put my
girls on this road… When their dad has time we’ll put the bikes into
the truck and leave our neighbourhood and go down to the Meewasin
Trail [dedicated walking/biking trail along the riverbank].
–Mother, urban low SES neighbourhood
Thus, low levels of neighbourhood-based physical activity do not
necessarily correspond to low levels of physical activity overall. The
ease of travel to other areas where opportunities for recreation are
found influences how likely families are to be active outside their
neighbourhood:
Saskatoon still isn’t such a big city that it’s difficult, I mean, you have
to plan your routes and think about where you’re going… to get there
in an efficient amount of time, but it’s still very achievable to get
around.
–Mother, urban middle SES neighbourhood
Neighbourhood social characteristics, amenities and role
modeling
Participants cited two key neighbourhood characteristics – safety
and recreational facilities – as influences on children’s activity, as
well as two social factors: the presence of other children in the
neighbourhood and parents’ own activity patterns.
The perceived safety of the neighbourhood, in terms of traffic,
crime or both, played a substantial role in whether parents allowed
their children to engage in outdoor activity. Children generally
shared their parents’ perceptions of whether or not their neighbourhood was safe, and their behavioural choices reflected this.
[Biking in my neighbourhood is dangerous because] there’s a lot of cars
there. There’s a lot of pawnshops around that area too. There’s not
really a lot of safe places for bikers to ride.
–Girl, aged 13 years, urban low SES neighbourhood
The kids around here are very active because there’re so many parks
around here and it’s a really nice neighbourhood…It’s one of the most
safe neighbourhoods, so I could walk outside, like really late at night.
–Girl, aged 11 years, urban middle SES neighbourhood
In some low SES neighbourhoods, parks and green spaces are perceived as sites of criminal activity and other misuse. This creates a
vicious cycle, in which parks become less used by families and children for active play and recreation, as they were intended, which
in turn increases the level of illicit activities.
Along with safety, the presence of recreational facilities in a
neighbourhood supports children’s activity. Neighbourhood amenities facilitate children’s activity because their proximity reduces the
time required to get to them, which is often in short supply for
families.
The availability of having that soccer centre right here was awesome…
Let’s say if I was working, and my husband works till 5:30, if he had
to worry about getting home and getting [my son] to a football game
on the other side of town, [it] might not have happened. You’d have to
do more arranging with other parents and that kind of thing.
Social influences, in terms of the presence of other children and
parents, were cited by many participants as important to children’s
activity. The presence of other children in the neighbourhood facilitates activity, both because parents consider it safer and because
children enjoy being active with others. Parents reported that they
were more willing to allow their children to play or travel outside
in groups than on their own. This could include siblings as well as
other children. Even children who were heavily restricted in their
independent travel were often allowed to travel or play with friends
or siblings. One mother commented of her 11-year-old daughter:
Do you know, there isn’t really anywhere I send her by herself. She’s
allowed to go with her brother to the neighbourhood park. She’s
allowed to go with her brother to friends’ houses on the street. On a
busy bright day, I’d let them take the underpass together if there was
a yard sale that they wanted to get to or something. But for the most
part, I don’t let her do a lot by herself.
–Mother, urban high SES neighbourhood
Thus, parental perceptions of safety and resulting restrictions,
regardless of actual risk, play a significant role in children’s opportunities to be active outdoors in their neighbourhood, and the presence of other children can mitigate these restrictions.
The children who attend local schools tend to have friends nearby, which facilitates afterschool and weekend activity. Most children are allowed to travel to friends’ houses, corner stores or parks
with one or more other children. Other than school, these are the
main destinations that children use active transportation to get to,
and children living in semi-suburban and suburban neighbourhoods generally have better access to them, in terms of both perceived safety and proximity.
Familiarity with neighbours increased parents’ perceptions of
neighbourhood safety. In neighbourhoods where parents knew
many of their neighbours and trusted them, children were more
likely to be encouraged to play outside the home. Having other
children nearby who they know and with whom they are allowed
to play makes it easier and more enjoyable for children to be active:
One of the nice things we liked about this neighbourhood is that in
this area, in this crescent, [my son] can just go out and go…You feel
like kids need to have a play date these days and it’s nice to just say,
“Go out and go find a friend.”
–Mother, semi-suburban middle SES neighbourhood
One 14-year-old girl underscored the importance of having
friends around to do things with outdoors:
When you’re outside you just want to go for a walk, if you go alone it’s
not really fun, you get bored easily and you’re just walking around and
then if you’re with friends you can just talk to them and walk around
or go and play a game that you can’t really, like, play football by yourself or go play basketball by yourself, so it’s not as fun as with a bunch
of people.
–Girl, semi-suburban middle SES neighbourhood
In fact, parents often commented that their children are unlikely to initiate physical activity on their own.
Parents also felt that being active themselves contributed to their
children’s likelihood of activity by providing a positive role model.
Physical activity was encouraged in both parents and children
when family time is spent engaging in physical activity together. In
neighbourhoods perceived to be unsafe, companionship was an
especially important facilitator in increasing neighbourhood-based
physical activity. A positive cycle of influence can operate within
families, with parents striving to be active so that their children
will be, too.
One of my motivators is just to be . . . the role model for my kids. Like
when I go out running, I always ask them if they want to come.
–Mother, semi-suburban medium SES neighbourhood
Thus, neighbourhood-based activity is most likely to occur when
1) parents are active themselves and encourage their children to be
active, 2) other children are present, 3) places for recreation are
nearby and 4) the neighbourhood is perceived to be safe.
We’re just active people… we’re active with our kids so that I think
that’s the biggest thing but… I mean it’s a decent neighbourhood; you
can go out and play in the park or do whatever.
DISCUSSION
Parents and children in this study underscored the importance of
safe environments for children’s physical activity: streets or paths
they can cycle on without feeling threatened, parks and green
spaces free of criminal activity, and neighbourhoods where people
know each other and children have friends to play with. Although
urban, grid-pattern neighbourhoods with a high density of destinations may in principle promote active transportation or walkability, the higher levels of crime and traffic danger that tend to exist
in these areas may hinder both leisure and utilitarian walking, as
well as cycling,25 especially for children. In our study, while adult
participants acknowledged that their own behaviour is influenced
by environmental factors consistent with the concept of walkability (e.g., proximity to commercial destinations, walking trails and
beautiful scenery), children’s patterns were different. Those living
in neighbourhoods with more commercial destinations were actually less likely to walk there, mostly because of the heavy vehicle
traffic in these areas and, in some cases, reduced personal safety.
Thus, what makes a neighbourhood “walkable” appears to be different for children than for adults, and children’s physical activity
may be more influenced by social factors, including their parents’
behaviour, and particularly safety, than the built environment.
US research has demonstrated that crime levels in neighbourhoods are negatively related to physical activity levels.26 Consistent
with our findings, other research has demonstrated that inner-city
neighbourhoods have higher obesity rates and lower levels of
neighbourhood-based physical activity than do suburban neighbourhoods.21 Regardless of actual crime levels, perceptions of safety have an impact on physical activity levels. US and Canadian
studies have demonstrated that parents’ perceptions of neighbourhood safety influence the type and level of their children’s activities, with the result that children residing in neighbourhoods
perceived to be unsafe are more likely to be overweight or obese.2729
Canadian studies have found that, among children, BMI increases in low SES neighbourhoods29 and that among adults the
perception of traffic danger in many low SES neighbourhoods is a
barrier to neighbourhood walking.9 Our study corroborates this
finding by showing that low SES neighbourhoods are perceived to
be unsafe and consequently deter neighbourhood-based physical
activity.
Active transportation for children mostly relates to their travel
between home and school, as well as to friends’ houses and parks.
In Saskatoon, while most neighbourhoods are home to at least one
elementary school, children may attend any elementary school in
the city. Many children in this sample living in urban neighbourhoods attended schools some distance away, preventing them from
walking or cycling to school. Further research examining the degree
to which proximity figures into parents’ choices regarding schools
for their children would be worthwhile. Parents who place a lower
value on their children being able to walk to school may be less
likely to encourage physical activity in any case. As we found, even
when it is feasible for children to walk or bike to school, other variables, such as a parent’s own transportation habits, enter into parents’ decision whether to drive them or require them to make their
own way.
Children’s activity is not, and should not be, limited to active
transportation. Not surprisingly, having places for children to play
and engage in sports and recreation that are easy and safe to get to
was seen as facilitating their activity. This includes both neighbourhood facilities, such as parks and biking paths that children
can go to on their own or with friends (if their parents allow them
to), and amenities that are outside the neighbourhood but can still
be reached quickly and easily by driving. Where neighbourhood
facilities are lacking, children are more dependent on their parents’
support to be active; for example, if neighbourhood streets are considered unsafe for cycling, parents may choose to drive their children to safer areas to go for a bike ride. While this works for some
families, not all children have parents who are willing and able to
take them to recreational facilities outside their neighbourhood.
Our findings regarding the importance of safety and easy access
to recreational opportunities have important implications for
health equity. Studies have shown that outside of school hours,
parks are the primary location in which play and physical activity
occur for low-SES children, who tend to have limited access to other
open spaces or recreational venues30 and are least likely to access
registered sports because of cost constraints.29 Children in low-
income families could thus benefit greatly from having free, easily
accessible recreational opportunities in the form of parks and associated programming within their neighbourhoods. However, in our
study, suburban and semi-suburban neighbourhoods were viewed
by participants as quieter and safer from traffic and crime for children than urban neighbourhoods, which are where low-income
families are more likely to find affordable housing.
Active transportation and reduced car use have numerous positive environmental and social benefits, and for this reason designing neighbourhoods to facilitate walking holds much merit. On the
surface, the findings of this study might be seen as endorsing suburban neighbourhood design as a way to promote activity in children, because such neighbourhoods provide safer places for
children to play and travel. In fact, facilitating activity in children
is more complex; it requires understanding the barriers to physical
activity in all neighbourhoods and finding ways to expand the
opportunities experienced by suburban children into urban neighbourhood settings. These could include measures to improve actual and perceived safety, such as providing more supervised
recreational activities through increased neighbourhood surveillance and organized group activities, and ensuring that parks in
urban neighbourhoods are up to date, well maintained and adequately lit. Even more fundamentally, broader initiatives need to be
envisioned to increase community cohesion, reduce social
inequities and promote neighbourhood safety by reducing the root
causes of crime and social disorder.
Limitations
In this qualitative sample, we found that children residing in urban
neighbourhoods attended schools outside their neighbourhoods,
and as a result, took part in less neighbourhood-based physical
activity. This finding may not be applicable to cities in which children living in urban neighbourhoods are more likely to attend their
neighbourhood school. An additional limitation of the study is that
the relatively small sample of participants in the qualitative component may have provided only a small number of perspectives.
Expanding the study to a larger group of participants representing
a larger range of neighbourhood types could possibly provide additional information.
REFERENCES
1.
2.
3.
4.
5.
6.
7.
8.
Gauvin L, Richard L, Craig CL, Spivock M, Riva M, Gagnon H, et al. From
walkability to active living potential: An “ecometric” validation study. Am J
Prev Med 2005;28(2S2):126-33.
Boone-Heinonen J, Gordon-Larsen P, Guilkey DK, Jacobs Jr DR, Popkin BM.
Environment and physical activity dynamics: The role of residential selfselection. Psychol Sport Exerc 2011;12:54-60.
Levine J, Frank LD. Transportation and land-use preferences and residents’
neighbourhood choices: The sufficiency of “smart growth” in the Atlanta
regions. Transportation 2006;34(2):255-74.
Berry TR, Spence JC, Blanchard CM, Cutumisu N, Edwards J, Selfridge G.
A longitudinal and cross-sectional examination of the relationship between
reasons for choosing a neighbourhood, physical activity and body mass
index. Int J Behav Nutr Phys Act 2010;7(57). doi: 10.1186/1479-5868-7-57.
Rodriguez DA, Khattak AJ, Evenson KR. Can new urbanism encourage physical activity? Comparing a new urbanist neighbourhood with conventional
suburbs. J Am Planning Assoc 2006;72(1):43-54.
Geller AL. Smart growth: A prescription for livable cities. Am J Public Health
2003;93(9):1410-15.
Pendola R, Gen S. BMI, auto use, and the urban environment in San Francisco. Health Place 2007;13:551-56.
Frank LD, Andresen M, Schmid TL. Obesity relationships with community
design, physical activity and time spent in cars. Am J Prev Med 2004;27(2):8796.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
Berry TR, Spence JC, Blanchard CC, Cutumisu N, Edwards JJ, Nykiforuk CC.
Changes in BMI over 6 years: The role of demographic and neighborhood
characteristics. Int J Obes 2010;34(8):1275-83.
Sallis JF, Glanz K. The role of built environments in physical activity, eating
and obesity in childhood. Future Child 2006;16(1):89-108.
Giles-Corti B, Kelty SF, Zubrick SR, Villanueva KP. Encouraging walking for
transport and physical activity in children and adolescents: How important
is the built environment? Sports Med 2009;39(12):995-1009.
Frank L, Kerr J, Chapman J, Sallis J. Urban form relationships with walk trip
frequency and distance among youth. Am J Health Promot 2007;21(4S):305-11.
Rahman T, Cushing RA, Jackson RJ. Contributions of built environment to
childhood obesity. Mount Sinai J Med 2011;78:49-57.
Van Dyck D, Cardon G, Deforche B, De Bourdeaudhuij I. Lower neighbourhood walkability and longer distances to school are related to physical activity in Belgian adolescents. Prev Med 2009;48:516-18.
Norman GJ, Nutter SK, Ryan S, Sallis JF, Calfas KJ, Patrick K. Community
design and access to recreational facilities as correlates of adolescent physical
activity and body-mass index. J Phys Act Health 2006;3(S1):S118-S128.
Leung CW, Gregorich SE, Laraia BA, Kushi LH, Yen IH. Measuring the neighbourhood environment: Associations with young girls’ energy intake and
expenditure in a cross-sectional study. Int J Behav Nutr Phys Act 2010;7:52-61.
Singh GK, Siahpush M, Kogan MD. Neighborhood socioeconomic conditions,
built environments, and childhood obesity. Health Affairs 2010;29(3):503-12.
Saarloos D, Kim J, Timmermans H. The built environment and health: Introducing individual space-time behaviour. Int J Environ Res Public Health
2009;6:1724-43.
Oakes M. The (mis)estimation of neighbourhood effects: Causal inference for
a practicable social epidemiology. Soc Sci Med 2004;59:1929-52.
Zick CD, Smith KR, Fan JX, Brown BB, Yamada I, Kowaleski-Jones L. Running
to the store? The relationship between neighbourhood environments and the
risk of obesity. Soc Sci Med 2009;69(10):1493-500.
Lopez RP, Hynes HP. Obesity, physical activity, and the urban environment:
Public health research needs. Environ Health 2006;5:25. doi:10.1186/1476069X-5-25.
Lincoln YS, Guba E. Naturalistic Inquiry. Beverly Hills, CA: Sage Publications,
1985.
Merriam SB. Qualitative Research and Case Study Applications in Education,
2nd ed. San Francisco, CA: Jossey-Bass Publishers, 1998.
City of Saskatoon, Neighbourhood profiles. Available at: http://www.saskatoon.ca/departments/community%20services/planningdevelopment/futuregrowth/demographicandhousingdata/pages/neighbourhoodprofiles.aspx
(Accessed May 15, 2011).
Gauvin L, Riva M, Barnett T, Richard L, Craig CL, Spivock M, et al. Association between neighborhood active living potential and walking. Am J
Epidemiol 2008;167(8):944-53.
Gordon-Larsen P, McMurray R, Popkin M. Determinants of adolescent physical activity and inactivity patterns. Pediatrics 2000;105(6):1-8.
Timperio A, Salmon J, Ball K. Evidence-based strategies to promote physical
activity among children, adolescents and young adults: Review and update.
J Sci Med Sport 2004;7(1):20-29.
Weir LA, Etelson D, Brand DA. Parents’ perceptions of neighbourhood safety
and children’s physical activity. Prev Med 2006;43(3):212-17.
Oliver LN, Hayes MV. Neighbourhood socio-economic status and the prevalence of overweight Canadian children and youth. Can J Public Health
2005;96(6):415-20.
Loukaitou-Sideris A, Sideris A. What brings children to the park? Analysis and
measurement of the variables affecting children’s use of parks. J Am Planning
Assoc 2010;76(1):89-107.
RÉSUMÉ
Objectifs : Jusqu’à maintenant, très peu d’études se sont penchées sur
le processus par lequel le design communautaire, le milieu bâti et
l’environnement social influent sur l’activité physique des gens, en
particulier les enfants. Les enquêtes qualitatives sont utiles pour explorer
les perceptions et la prise de décisions, et pour comprendre les processus
en jeu dans les interactions des gens avec leur environnement. Notre
étude fait appel à des méthodes qualitatives pour approfondir la
compréhension des liens entre l’environnement du quartier et le profil
d’activité des enfants.
Méthode : Des données ont été recueillies à la faveur d’entretiens semidirigés auprès de 24 dyades parents-enfants (les enfants ayant de 10 à 14
ans). Les familles habitaient des quartiers au revenu médian variable (du
plus faible au plus élevé) et qui représentaient les trois grands types de
design observés à Saskatoon : urbain, semi-suburbain et suburbain.
Résultats : Parents et enfants ont souligné l’importance que
l’environnement soit sûr pour l’activité physique des enfants : des rues ou
des sentiers où l’on peut faire de la bicyclette sans se sentir menacé, des
parcs et des espaces verts sans activités criminelles, et des quartiers où les
gens se connaissent et où les enfants ont des camarades avec qui jouer.
Bien que les quartiers urbains aux rues quadrillées, denses en points
d’intérêt, favorisent en principe le transport actif, les taux de criminalité
plus élevés et les dangers de la circulation qui ont tendance à exister dans
ces quartiers peuvent entraver l’activité physique des enfants.
Conclusion : Tenter de comprendre ce qui facilite l’activité chez les
enfants est une tâche complexe. Elle exige de connaître les obstacles à
l’activité physique présents à l’échelle des quartiers ainsi que les facteurs
sociaux et perceptuels qui agissent de façon interdépendante pour
favoriser ou entraver l’activité physique des enfants.
Mots clés : milieu bâti du quartier; enfants; méthode qualitative;
sécurité; activité physique
There’s No Such Thing as Bad Weather, Just the Wrong Clothing:
Climate, Weather and Active School Transportation in Toronto, Canada
Raktim Mitra, PhD,1 Guy Faulkner, PhD2
ABSTRACT
Objective: Climatic conditions may enable or deter active school transportation in many North American cities, but the topic remains largely
overlooked in the existing literature. This study explores the effect of seasonal climate (i.e., fall versus winter) and weekly weather conditions
(i.e., temperature, precipitation) on active travelling to school across different built and policy environments.
Methods: Home-to-school trips by 11-12-year-old children in the City of Toronto were examined using data from the 2006 Transportation Tomorrow
Survey. Binomial logistic regressions were estimated to explore the correlates of the choice of active (i.e., walking) versus non-active (i.e., private
automobile, transit and school bus) mode for school trips.
Results: Climate and weather-related variables were not associated with choice of school travel mode. Children living within the sidewalk snow-plough
zone (i.e., in the inner-suburban neighbourhoods) were less likely to walk to school than children living outside of the zone (i.e., in the inner-city
neighbourhoods).
Conclusion: Given that seasonality and short-term weather conditions appear not to limit active school transportation in general, built environment
interventions designed to facilitate active travel could have benefits that spill over across the entire year rather than being limited to a particular season.
Educational campaigns with strategies for making the trip fun and ensuring that the appropriate clothing choices are made are also warranted in
complementing built environment modifications.
Key words: Climate; weather; school travel; walking; built environment
C
hildren who walk or cycle for school transportation tend to
be more physically active overall than those who use other
travel modes.1 Policy and program initiatives in North America have focused on active school transportation (AST), i.e., walking and cycling, as a means to increase physical activity levels
among children and youth, and in the longer term, to reverse the
current trend of increasing obesity rates. This policy concern is
matched by an emerging literature that has explored the barriers
and correlates of active school transportation.2 Much of the recent
literature considers the potential influence of travel distance and
the neighbourhood environment on AST uptake.3-6
For many North American cities that have marked seasonal variation in climate, weather conditions are popularly identified as
enablers or barriers to participation in outdoor physical activity,
including AST. For example, a study that analyzed data from a consumer styles survey in the US revealed that 18.6% of parents identified bad weather as a barrier to their children (5-18 years) walking
to or from school at least once per week.7 In Canada, parents who
continued to drive their children to school after a School Travel
Planning intervention most often cited weather (21.0%) as the reason for doing so.8
However, it is not clear from such North American evidence
whether the seasonal climate (e.g., summer, fall and winter) and
short-term weather conditions related to climatic variations
(e.g., temperature, rain, snow) do influence decision-making behaviour
about school travel or whether perhaps these natural conditions are
used to externalize the reason for driving. The actual effects of seasonal climate and objectively measured weather conditions associated with these seasonal climates remain largely overlooked in the
school transportation literature. Existing evidence on the association between climatic conditions and active travel is limited and
inconclusive. For example, Sirard et al.9 studied rates of active travel among students going to eight elementary schools in Columbia,
South Carolina, US, and found no association between weather conditions, temperature and the number of students actively travelling
(walking and cycling) to school. Robertson-Wilson et al.10 studied
students from 76 high schools in Ontario; season of the year (summer, fall and winter), average weekly temperature and weekly precipitation were not associated with AST. In Norway, where there are
extreme differences in weather conditions between the summer and
winter months, the number of children walking to and from school
actually increases in winter (24% in the fall to 65% in the winter) as
a result of a decline in cycling during that season.11
Author Affiliations
1. Department of Geography and Program in Planning, University of Toronto,
Toronto, ON
2. Faculty of Kinesiology and Physical Education, University of Toronto, Toronto, ON
Correspondence: Raktim Mitra, School of Urban and Regional Planning, Ryerson
University, 105 Bond St., Toronto, ON M5B 1Y3, E-mail: [email protected].
Acknowledgements: This research was funded by the Built Environment, Obesity
and Health Strategic Initiative of the Heart and Stroke Foundation and the Canadian
Institutes of Health Research. The first author wishes to acknowledge support from the
Natural Sciences and Engineering Research Council, Canada.
CLIMATE AND ACTIVE SCHOOL TRAVEL
Figure 1.
Study area – City of Toronto
Kilometres
Data: Environment Canada,14 City of Toronto.15
Seasonal climate and weather conditions may have received less
attention in the current literature because these natural conditions
are non-modifiable. However, there are two important reasons why
their effects on school travel outcome should be explored. First, if
climate and weather conditions have an important influence on
school travel behaviour, then expensive public investments leading
to improvements in the built environment can only produce limited benefits, perhaps only by improving walking rates in months
when weather conditions are generally good. This should be an
important consideration before making such investments. Second, if
climate and weather conditions do not have an important influence,
then there is a need for more focused interventions that target households situated within a walkable distance from school throughout
the school year and that are tailored appropriately for each season.
Within this context, this study explored the effects of seasonal
climate and objectively measured weather conditions on children’s
AST uptake in the City of Toronto. The potential role of snow-clearing
policy was also taken into account. To our knowledge, this is the
first study to rigorously examine climate-related effects on children’s school travel mode in a large North American city with
marked seasonal variation in climatic conditions.
Study design
The school transportation behaviour of children 11-12 years old in
the City of Toronto was examined. The primary hypothesis of the
paper was that once the variations in household socio-economic
characteristics and the travel distance are taken into account, the
effect of seasonal climate and related weather conditions on active
school travel, particularly walking, would be minimal.
ture in Toronto was -2.4oC (minimum temperature recorded in
2006 was -13.1oC). In contrast, the mean temperatures in July and
October were 23.4oC and 9.7oC respectively.14 Total precipitation in
2006 was 957.1 mm, which occurred mostly in the summer and
winter months. The city also has different snow management policies for its inner-city and inner-suburban neighbourhoods, thanks
to the legacy of a political and geographic amalgamation in 1998
(Figure 1). The 2000 Snow Plan requires owners of inner-city properties to clear sidewalks of snow and ice within 12 hours after a
snowfall. In contrast, inner-suburban residents of Toronto continue to enjoy publicly funded, mechanical sidewalk clearing services,
as they did before the amalgamation.15
Travel data
This study uses school travel data from the 2006 Transportation
Tomorrow Survey (TTS). The TTS is a repeated cross-sectional survey of travel behaviour in southern Ontario and covers a 5% sample of all households in the study area.16 The 2006 TTS data were
collected for a randomly selected weekday in fall or winter; an adult
household member proxy reported travel data (e.g., origin/destination of trips, trip start time, purpose, primary travel mode) for all
trips by household members aged 11 years and older, associated
with the day before the interview. For this study, all home-to-school
trips between the 6h00-9h30 time interval (n=2,520) were extracted. Only the students who travelled to public and Catholic schools
were included; these students are expected to attend schools that
are closest to their residential locations. The school trips were made
between September 2006 and January 2007. School travel mode
was determined on the basis of the proxy reported data on primary
mode of travel.
Study area
Toronto is the largest city in Canada, with a population of over
2.5 million.12 The city is situated in a continental climate zone (Köppen
climate classification: Dfa).13 In 2006, the mean February temperaS36 REVUE CANADIENNE DE SANTÉ PUBLIQUE • VOL. 103, SUPPLÉMENT 3
Socio-economic and travel distance data
Household socio-demographic information and the distance
between home and school locations were used as control variables
Table 1.
Description of Variables and Summary Statistics
%
(n=1992)
Control Variables
Age: The child was 12 years old (reference: 11 years).
Sex: The child was a male (reference: female).
Number of Children: Number of school-age children below driving age (4-15 years) in the household.
Single-adult Household: There was only one adult household member (>17 yrs) in the household
(reference: >1 adults in a household).
Vehicles per Licensed Driver: Number of vehicles in the household per licensed driver.
Distance: Minimum travel distance (i.e., shortest path using a street network) between home and school (km).
Low-income Neighbourhood:* The child’s residential neighbourhood was a low-income neighbourhood,
i.e., median household income was <CAD $39,400. Median household income was estimated by taking a median of the
census dissemination area median household incomes (reference: not a low-income neighbourhood).
Climate and Weather Variables
Seasonal Climate:† (reference: Fall AND outside of the sidewalk snow-plough zone).
Winter AND outside of the sidewalk snow-plough zone.
Fall AND inside of the sidewalk snow-plough zone.
Winter AND inside of the sidewalk snow-plough zone.
Temperature (6-9:30 am):‡ Weekly average of hourly temperatures (°C) between the 06h00-09h30 time interval.
Max Temperature: Weekly average of daily maximum temperatures (°C).‡
Min Temperature: Weekly average of daily minimum temperatures (°C).‡
Mean Temperature: Weekly average of daily mean temperatures (°C).‡
Precipitation Days:‡ Number of days in a week with precipitation.
Average Precipitation:‡ Average precipitation in a week (mm).‡
Snow Days:‡ (reference: residence located outside of the sidewalk snow-plough zone AND no snow during the week).
Residence located outside of the sidewalk snow-plough zone AND ≥1 days of snow during the week.
Residence located within the sidewalk snow-plough zone AND no snow during the week.
Residence located within the sidewalk snow-plough zone AND ≥1 days of snow during the week.
Mean (SD)
52.2
52.8
1.92 (0.84)
12.4
0.72 (0.42)
1.05 (0.69)
16.3
8.9
54.8
12.9
5.08 (5.64)
9.70 (5.91)
2.52 (5.66)
6.13 (5.73)
2.16 (1.20)
2.30 (3.11)
9.1
58.7
8.9
Note: Variables in italics were excluded from the multivariate logistic regression specifications.
* Individual household income data were not available. Median household income was calculated using 2006 population census data from Statistics Canada.
Average household size for the sample was 4.3 (SD=1.28). In a large metropolitan area such as Toronto (i.e., population >500,000), the low income cut-off,
defined by Statistics Canada, was CAD $39,399 for a four-member household.17
† Fall: Travel data were collected between September 10 and December 17, 2006; Winter: Travel data were collected in January 2007.
‡ Weather data were averaged/aggregated for weekdays only; holidays and weekends were excluded.14
(Table 1). The socio-demographic data were taken from the 2006
TTS. However, the TTS did not provide information on household
income. As a result, the median household income data from the
population census of Canada were aggregated to identify the lowincome neighbourhoods. Shortest path network distance (i.e., minimum travel distance) between home and school locations of a
child was computed using Toronto’s street network file (DMTI CanMap© RouteLogistics file, version 2007.3).
Climate and weather data
School trip data were collected between September 2006 and January 2007. These data were categorized into fall (78.2% of total trips;
data collected between September 10 and December 17, 2006) and
winter (21.8% of total trips; data collected in January 2007) trips,
in order to capture potentially climatically moderated differences in
travel behaviour. In addition, seven objectively measured weatherrelated variables were computed for three weather stations located
inside and near the City of Toronto (Toronto City, Toronto Lester
B. Pearson Airport and Toronto Buttonville Airport), using data
from Environment Canada customized climate search (Figure 1).14
Weekly average for each weather variable was used, computed from
an average of all school days in the week of data collection (Table
1). The 2006 TTS did not provide information on the date of travel. For each household, the nearest weather station (straight line
distance) was identified to determine the weekly weather condition for that household. In order to account for Toronto’s snow
management policy, households located inside and outside of the
sidewalk snow-plough zone were identified.
Statistical analysis
Binomial logistic regressions were estimated to explore the correlates of active (i.e., walking) versus non-active (i.e., private auto-
mobile, public transit or school bus) mode choice for school trips.
The effect of weather on walking and cycling are potentially different; these two modes also have different costs of access
(i.e., equipment, safety) and infrastructure requirements. In our sample,
only 16 children cycled to school and, because of this, cycling trips
were excluded from further statistical analysis. In addition, we
assumed that beyond a distance of 3.2 km (2 miles) the transportation choices become primarily limited to motorized modes.18
As a result, this study only explored the travel behaviour of children
living within 3.2 km from their schools. Adjusting for missing data
and outliers, the final dataset included 1,992 home-to-school trips.
The results are presented in terms of odds ratios (OR = exp(βî)),
which demonstrates the relation between a variable i and the odds
of active travel. Statistical analyses were performed with R© version
2.9.0.
The bivariate association between active travel and each of the
climate and weather-related variables was first examined; effect
plots were generated for the variables that demonstrated statistically significant associations. The degree of multi-collinearity across
the climate and weather variables was also explored before the
multivariate analysis. Not surprisingly, weather conditions (e.g., temperature, precipitation) were highly correlated with seasonal climate (i.e., fall and winter). To overcome this collinearity problem,
separate models were estimated to explore the effects of seasonal
climate and weekly weather conditions. Some of the weather variables were also highly correlated (r>0.9). Those correlated variables
(e.g., Temperature 6-9:30 am, Max temperature, Min temperature,
Mean temperature; and Precipitation days, Average precipitation)
were entered into the multivariate (i.e., adjusted) model one at a
time, and only the ones with strongest association (e.g., Temperature 6-9:30 am; Precipitation days) were included in the final model
specification.
Figure 2.
Rates of walking and car (i.e., private automobile) trips by distance
Table 2.
RESULTS
Logistic regression models of mode choice were estimated using
travel data for 1,992 students aged 11-12 years. The majority
(62.7%) walked to school; the other 37.3% travelled by private
automobile, school bus or transit (Table 2). Most students (79.8%)
lived within 1.6 km (1 mile) of their schools. Of these students,
72.7% walked to school, 17.6% were driven, and 9.7% used school
bus or transit as their travel mode.
Table 2 also demonstrates that school transportation mode shares
(percentage of total school trips) were not different between fall
and winter seasons, suggesting that seasonal climate does not have
an influence on mode choice. The rates of walking and car (i.e., private automobile) trips at different travel distances were also compared (Figure 2). Figure 2 indicates that in both fall and winter,
more children (with or without caregivers) walked than were driven to school when the household locations were within 1.4 km of
school. Not surprisingly, walking rates declined with increased distance, but no systematic variation across seasons was evident. This
result suggests that the perception of a “walkable distance” was not
different between fall and winter seasons. The model results further confirmed these observations (Table 3). Children in general
were less likely to walk within the sidewalk snow-plough zone
(i.e., in the inner-suburban neighbourhoods) compared with the areas
outside of the zone (i.e., in the inner-city neighbourhoods); seasonal climate (i.e., fall versus winter conditions) was not associated with the likelihood of walking within any of these urban
locations.
With regard to weekly weather conditions, the probability of
walking was inversely correlated (bivariate) with the number of
weekdays with precipitation and the average weekly precipitation
(mm) (Figure 3). There was also an association between walking
and snow days across the two snow-plough zones. However, most
of these correlations disappeared in the multivariate (i.e., adjusted)
model. Table 3 suggests that objective measures of weather were
not associated with mode choice when household socio-economic
characteristics and school travel distance were taken into account.
School Travel Mode Share (Percentage of Total
School Trips)
All Seasons
Fall
(n=1992) (n=1558)
Walk
Private automobile
School bus
Transit
62.7
20.5
9.6
7.2
62.6
20.7
9.5
7.2
Winter
(n=434)
62.9
19.8
9.9
7.4
Chi-square*
0.06
(p=0.807)
* Difference in mode choice distribution between fall and winter seasons;
calculated on the basis of the total number of trips.
The only statistically significant variable was “Snow Days”. However, the observed effect relates entirely to the geographic location
of the households, similar to our earlier observation related to seasonal climate. The amount of snowfall was not associated with the
likelihood of walking.
In addition to children living outside of the sidewalk snowplough zone (i.e., in the inner-city neighbourhoods), older students
and males were more likely to walk to school than the younger students and females (Table 3). The propensity of walking decreased
with an increased distance between home and school locations,
and with an increased number of vehicles in the households per
licensed driver.
DISCUSSION
Our findings indicate that seasonal climate and weekly weather
conditions do not appear to be major influencing factors on choice
of school travel mode in Toronto. Distance between home and
school remained the strongest correlate of travel mode, supporting
what has been reported elsewhere.19,20 Students living in innersuburban neighbourhoods walked to school less. Even with the
removal of a potential barrier to walking in the winter in the innersuburban neighbourhoods (i.e., sidewalk snow removal by the City
of Toronto), students were still less likely to walk, suggesting consistency in travel mode regardless of season. While the city-run,
sidewalk snow plough service remains a political legacy in Toronto’s inner-suburban neighbourhoods, this policy does not facilitate
children’s AST uptake.
Table 3.
Correlates of Walking to School
Base Model
OR (95% CI)
Age
11 years
12 years
Sex
Female
Male
Number of children
Single-adult household
Vehicles per licensed driver
Distance
Low-income neighbourhood
No
Yes
Seasonal climate
Fall AND outside of the sidewalk snow-plough zone
Winter AND outside of the sidewalk snow-plough zone
Fall AND inside of the sidewalk snow-plough zone
Winter AND inside of the sidewalk snow-plough zone
Precipitation days
Temperature (6-9:30 am)
Precipitation days × Temperature (6-9:30 am)
Snow days
Outside of the sidewalk snow-plough zone AND
no snow during the week
Outside of the sidewalk snow-plough zone AND
≥1 days of snow during the week
Within the sidewalk snow-plough zone AND
no snow during the week
Within the sidewalk snow-plough zone AND
≥1 days of snow during the week
Intercept
–2 [L(0) – L(B)]
McFadden’s ρ2 (adjusted)
AIC†
p
Seasonal Climate Model
OR (95% CI)
p
Weather Conditions Model
OR (95% CI)
p
1.00
1.60 (1.28-1.99)
0.000
1.00
1.58 (1.26-1.97)
0.000
1.00
1.58 (1.26-1.97)
0.000
1.00
1.34 (1.08-1.67)
1.08 (0.95-1.24)
1.07 (0.77-1.50)
0.62 (0.48-0.81)
0.13 (0.11-0.16)
0.008
0.241
0.681
0.000
0.000
1.00
1.38 (1.11-1.72)
1.09 (0.95-1.25)
1.03 (0.74-1.44)
0.64 (0.49-0.83)
0.13 (0.10-0.15)
0.004
0.208
0.865
0.000
0.000
1.00
1.37 (1.10-1.70)
1.09 (0.95-1.25)
1.03 (0.73-1.43)
0.65 (0.50-0.85)
0.13 (0.11-0.16)
0.005
0.209
0.879
0.001
0.000
1.00
1.24 (0.91-1.68)
0.166
1.00
1.23 (0.91-1.67)
0.159
1.00
1.21 (0.89-1.65)
0.213
1.00
0.73 (0.46-1.16)
0.57 (0.43-0.75)
0.56 (0.38-0.82)
0.183
0.000
0.003
0.94 (0.84-1.07)
1.03 (0.97-1.09)
0.99 (0.97-.01)
0.362
0.342
0.441
1.00
1.06 (0.62-1.82)
0.828
0.64 (0.48-0.86)
0.003
0.71 (0.42-1.20)
0.203
12.18 (7.93-18.72) 0.000 17.82 (10.99-28.88) 0.000 16.94 (9.64-29.75) 0.000
646.6
964.0
665.9
0.24 (0.24)
0.25 (0.25)
0.25 (0.25)
2002.0
1990.6
1994.7
Note: Variables in bold are significant at α=0.01.
OR= odds ratios, CI=confidence interval.
† Akaike information criterion.
Figure 3.
Effect of weekly weather on walk to school (bivariate logistic regression results)
Note: The dotted lines represent the 95% percentile brackets around the fitted lines.
* Snow Days: A = Outside the sidewalk snow-plough zone and no snow; B = Outside the sidewalk snow-plough zone and ≥1 days of snow during the week;
C = Inside the sidewalk snow-plough zone and no snow; D = Inside the sidewalk snow-plough zone and ≥1 days of snow during the week.
In this study, weather variables were based on weekly averages
rather than the weather on the day of the school trip. In other
words, the variables explored in this study capture the prevalent
weekly weather conditions and their potential contribution to the
formation of individual travel patterns. We could not examine
responses to extreme weather conditions on any given day. Extreme
weather may produce short-term changes in mode choice anywhere on the planet for that matter, and for any activity. However,
given the utilitarian necessity of the school trip, it may be that
choice of transportation modes for travelling to elementary schools
is largely habitual in nature and less influenced by external factors
such as the climate and weather unless in those extreme conditions
where walking becomes dangerous or extremely difficult. Developed from a qualitative study of parent/child interviews focused
on AST in Toronto, a two-stage AST decision-making process was
described by Faulkner et al.,21 one decision being whether or not the
child needs escorting to/from school and a second decision about
mode choice, which is largely influenced by perceptions as to the
easiest and most convenient way to travel. Importantly, these initial decisions had developed into routinized behaviour that no
longer required a conscious decision-making process. For example,
for the parents of children who walked to school, walking remained
the most convenient option even in the winter months.21 Additionally, others have reported that adults who enjoy exercise22 or
who walk for any reasons in and around their neighbourhoods23
are less likely to mention weather as a barrier. Overall, this again
points to the potential stability of choice of school travel mode
based on preferences and perceptions of convenience rather than
weather per se.
With more extensive sampling during winter and a more rigorous
exploration of climate and weather conditions at different travel
distances and across different built and policy environments, this
study significantly extends existing research on the relation of climate and weather conditions with school travel. However, further
research is required to confirm the absence of an association given
some limitations of this study. First, there may be some variability
between weather conditions at the weather stations and those conditions in close vicinity to the household, although the majority of
households were within 10 km of a weather station (mean 8.96±4.03
km). Second, our findings also require replication in other geographic locations in Canada where there is more extreme seasonal
variation. Yet local policy should be based upon local evidence, and
our results have important implications for promoting active school
travel within the Greater Toronto Area, Canada’s largest and most
culturally diverse metropolitan region. In terms of policy, our results
suggest that climate and weather conditions do not have any significant effect on AST in this study location. The results are encouraging for policies directed toward improvements in the built
environment to increase rates of walking and physical activity
among children. Given that seasonality and short-term weather conditions appear not to limit AST uptake in general, the potentially
enormous capital cost of built environment interventions designed
to facilitate AST could have benefits that spill over across the entire
year rather than being limited to a particular season.
In terms of practice, there is growing interest in initiatives targeting drivers of children from households located within a walkable distance from school.8,24 Our findings suggest that seasonal
climate and weather should not be actual barriers to AST in Toronto and, moreover, that parents are likely driving their children to
school even when the climate is most conducive to supporting AST
(e.g., in the spring and fall). For example, in the fall of 2006
(between September and December), 17.6% of all students who
lived within 1.6 km of their schools were driven to school, and in
the winter of 2007, 17.5% were driven. Educational campaigns to
promote winter walking with strategies for overcoming barriers and
parental safety concerns, making the trip fun, and ensuring that
clothing choices are appropriate, are also warranted in complementing built environment modifications that facilitate active
school travel.
REFERENCES
1.
Faulkner GEJ, Buliung RN, Flora PK, Fusco C. Active school transport, physical activity levels and body weight of children and youth: A systematic review.
Prev Med 2009;48(1):3-8.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
Stewart O. Findings from research on active transportation to school and
implications for Safe Routes to School programs. J Planning Literature
2011;26(2):127-50.
McDonald NC. Children’s mode choice for the school trip: The role of distance and school location in walking to school. Transportation 2008;35(1):2335.
McMillan TE. The relative influence of urban form on a child’s travel mode
to school. Transportation Research Part A 2007;41(1):69-79.
Mitra R, Buliung RN, Roorda MJ. Built environment and school travel mode
choice in Toronto, Canada. Transportation Research Record 2010;2156:150-59.
Panter JR, Jones AP, van Sluijs EMF, Griffin SJ. Neighborhood, route, and
school environments and children’s active commuting. Am J Prev Med
2010;38(3):268-78.
Martin S, Carlson S. Barriers to children walking to or from school: United
States, 2004. MMWR 2005;54(38):949-52.
Buliung R, Faulkner GEJ, Beesley T, Kennedy J. School travel planning: Mobilizing school and community resources to encourage active school transportation. J School Health 2011;81:704-12.
Sirard JR, Ainsworth BE, McIver KL, Pate RR. Prevalence of active commuting
at urban and suburban elementary schools in Columbia, SC. Am J Public
Health 2005;95(2):236-37.
Robertson-Wilson JE, Leatherdale ST, Wong SL. Social-ecological correlates of
active commuting to school among high school students. J Adolesc Health
2008;42(5):486-95.
Borrestad LAB, Andersen LB, Bere E. Seasonal and socio-demographic determinants of school commuting. Prev Med 2011;52:133-35.
Statistics Canada. Population and Dwelling Counts, for Canada, Provinces
and Territories, and Census Divisions, 2006 and 2001 Censuses – 100% Data,
2008. Available at: http://www12.statcan.ca/english/census06/data/popdwell/
Table.cfm?T=702&PR=35&S=0&O=A&RPP=25 (Accessed July 5, 2012).
McKnight TL, Hess D. Physical Geography: A Landscape Appreciation, 7th ed.
Upper Saddle River, NJ: Prentice Hall, 2002;198-239.
Environment Canada. National Climate Data and Information Archive, 2011.
Available at: http://climate.weatheroffice.gc.ca (Accessed November 25, 2011).
City of Toronto. Snow – City’s First Snow Plan 2000. 2011. Available at:
http://www.toronto.ca/transportation/snow/plan.htm (Accessed November
25, 2011).
Data Management Group. Transportation Tomorrow Survey, 2008. University of Toronto. Available at: http://www.dmg.utoronto.ca/transportationtomorrowsurvey/index.html (Accessed July 2, 2010).
Statistics Canada. Low income cut-offs for 2006 and low income measures for
2005. 2010. Available at: http://www.statcan.gc.ca/bsolc/olc-cel/olccel?lang=eng&catno=75F0002M2007004 (Accessed August 26, 2012).
Nelson NM, Foley E, O’Gorman DJ, Moyna NM, Woods CB. Active commuting to school: How far is too far? Int J Behav Nutr Phys Act 2008;5:1.
Mitra R, Buliung RN. Built environment correlates of active school transportation: Neighborhood and the modifiable areal unit problem. J Transport
Geography 2012;20(Special Issue on Child and Youth Mobility):51-61.
Wong B, Faulkner GEJ, Buliung RN. GIS measured environmental correlates
of active school transport: A systematic review of 14 studies. Int J Behav Nutr
Phys Act 2011;8:39.
Faulkner GEJ, Richichi V, Buliung R, Fusco C, Moola F. What’s “Quickest and
Easiest?”: Parental decision making about school trip mode. Int J Behav Nutr
Phys Act 2010;7:62.
Salmon J, Owen N, Crawford D, Bauman A, Sallis JF. Physical activity and
sedentary behaviour: A population-based study of barriers, enjoyment, and
preference. Health Psychol 2003;22(2):178-88.
Humpel N, Owen N, Iverson D, Leslie E, Bauman A. Perceived environment
attributes, residential location, and walking for particular purposes. Am J Prev
Med 2004;26(2):119-25.
McDonald NC, Aalborg AE. Why parents drive children to school: Implications for safe routes to school programs. J Am Planning Assoc 2009;75(3):33142.
RÉSUMÉ
Objectif : Les conditions climatiques pourraient favoriser ou entraver le
transport scolaire actif dans de nombreuses villes d’Amérique du Nord,
mais ce sujet est en grande partie inexploré dans la recherche existante.
Notre étude porte sur l’effet du climat saisonnier (p. ex., automnal ou
hivernal) et des conditions atmosphériques hebdomadaires (température,
précipitations) sur le transport scolaire actif dans différents milieux bâtis
et environnements politiques.
Méthode : Nous avons examiné les trajets de la maison à l’école
d’enfants de 11 et 12 ans vivant à Toronto à l’aide des données du
Sondage pour le système de transports de demain de 2006. Des
régressions logistiques binomiales ont été estimées pour explorer les
corrélats du choix d’un mode de transport actif (comme la marche) ou
non actif (voiture privée, transports en commun, autobus scolaire) pour
se rendre à l’école.
Résultats : Les variables liées au climat et à la météo n’étaient pas
associées au choix du mode de transport scolaire. Les enfants vivant à
l’intérieur de la zone de déneigement des trottoirs (c.-à-d. dans les
quartiers de la proche banlieue) étaient moins susceptibles de se rendre à
l’école en marchant que les enfants vivant hors de cette zone (c.-à-d.
dans les quartiers du centre-ville).
Conclusion : Étant donné que les cycles saisonniers et les conditions
atmosphériques de courte durée ne semblent pas limiter le transport
scolaire actif en général, les interventions sur le milieu bâti conçues pour
faciliter les déplacements actifs pourraient avoir des répercussions
positives toute l’année plutôt que de se limiter à une saison particulière.
On devrait aussi envisager, pour compléter des modifications au milieu
bâti, des campagnes de sensibilisation comportant des stratégies pour
rendre le trajet amusant et pour aider les enfants à choisir des vêtements
appropriés.
Mots clés : climat; temps; transport scolaire; marche; milieu bâti
Safe Cycling: How Do Risk Perceptions Compare With Observed Risk?
Meghan Winters, PhD,1 Shelina Babul, PhD,2 H.J.E.H. (Jack) Becker, MBA,3 Jeffrey R. Brubacher, MD, MSc,4
Mary Chipman, PhD,5 Peter Cripton, PhD,6 Michael D. Cusimano, MD, PhD,7 Steven M. Friedman, MD,8
M. Anne Harris, PhD,9 Garth Hunte, MD,4 Melody Monro, MPA,10 Conor C.O. Reynolds, PhD,11
Hui Shen, MSc,10 Kay Teschke, PhD10
ABSTRACT
Objective: Safety concerns deter cycling. The Bicyclists’ Injuries and the Cycling Environment (BICE) study quantified the injury risk associated with
14 route types, from off-road paths to major streets. However, when it comes to injury risk, there may be discordance between empirical evidence and
perceptions. If so, even if protective infrastructure is built people may not feel safe enough to cycle. This paper reports on the relationship between
perceived and observed injury risk.
Methods: The BICE study is a case-crossover study that recruited 690 injured adult cyclists who visited emergency departments in Toronto and
Vancouver. Observed risk was calculated by comparing route types at the injury sites with those at randomly selected control sites along the same route.
The perceived risk was the mean response of study participants to the question “How safe do you think this site was for cyclists on that trip?”, with
responses scored from +1 (very safe) to -1 (very dangerous). Perceived risk scores were only calculated for non-injury control sites, to reduce bias by the
injury event.
Results: The route type with the greatest perceived risk was major streets with shared lanes and no parked cars (mean score = -0.21, 95% confidence
interval [CI]: -0.54-0.11), followed by major streets without bicycle infrastructure (-0.07, CI -0.14-0.00). The safest perceived routes were paved multiuse paths (0.66, CI 0.43-0.89), residential streets (0.44, CI 0.37-0.51), bike paths (0.42, CI 0.25-0.60) and residential streets marked as bike routes with
traffic calming (0.41, CI 0.32-0.51). Most route types that were perceived as higher risk were found to be so in our injury study; similarly, most route
types perceived as safer were also found to be so. Discrepancies were observed for cycle tracks (perceived as less safe than observed) and for multiuse
paths (perceived as safer than observed).
Conclusions: Route choices and decisions to cycle are affected by perceptions of safety, and we found that perceptions usually corresponded with
observed safety. However, perceptions about certain separated route types did not align well. Education programs and social media may be ways to
ensure that public perceptions of route safety reflect the evidence.
Key terms: Safety; transportation; injury; environmental design
L
ack of safety is a major deterrent to cycling. Cyclists and potential cyclists report safety concerns related to motor vehicle traffic and poor weather.1,2 These concerns are valid: on a per-trip
basis, cycling is more dangerous than car travel, with US data showing that the fatality rate per bike trip was about 2.3 times higher
than that for automobile trips and that the police-reported injury
rate per bike trip was about 1.8 times higher than that for automobile trips.3 Similarly, the 73 cyclists killed in Canada in 2006 represented 2.5% of all traffic-related deaths,4 although cycling
represents only 1.3% of commuter travel.5 Of course, motor vehicles are not the only cause of bicycle crashes. Observational studies indicate that a large proportion of injuries and conflicts result
from falls and from collisions with route infrastructure (streetcar
tracks, curbs, potholes, etc.) and sometimes with pedestrians, other
cyclists and animals.6-8 These incidents are less likely to be reported to police or captured in insurance records.7,9,10
While the bulk of cycling safety research focuses on individualbased protection (e.g., helmet usage), population-level strategies
related to the built environment and the provision of cycling infrastructure hold promise to both prevent crashes and encourage
cycling.11 Worldwide, purpose-built bicycle-only facilities (e.g., bike
Author Affiliations
1.
2.
3.
4.
Faculty of Health Sciences, Simon Fraser University, Vancouver, BC
Department of Pediatrics, University of British Columbia, Vancouver, BC
Third Wave Cycling Group Inc., BC
Department of Emergency Medicine, Faculty of Medicine, University of British
Columbia, BC
5. Dalla Lana School of Public Health, University of Toronto, Toronto, ON
6. Department of Mechanical Engineering, University of British Columbia, BC
7. Neurosurgery, St. Michael’s Hospital, University of Toronto, ON
8. Faculty of Medicine, University of Toronto, University Health Network, Toronto,
ON
9. Occupational Cancer Research Centre, Cancer Care Ontario
10. School of Population and Public Health, University of British Columbia, BC
11. Liu Institute for Global Issues, University of British Columbia
Correspondence: Meghan Winters, Faculty of Health Sciences, Simon Fraser
University, Blusson Hall Rm 11522, 8888 University Dr., Burnaby, BC V5A 1S6,
Tel: 778-782-9325, Fax: 778-782-5927, E-mail: [email protected]
Canada and the Canadian Institutes of Health Research (Institute of Musculoskeletal
Health and Arthritis, and Institute of Nutrition, Metabolism and Diabetes).
Jeffrey R. Brubacher is a Michael Smith Foundation for Health Research Scholar.
Michael D. Cusimano is funded by the Canadian Institutes of Health Research and
the Ontario Neurotrauma Foundation. We thank all the study participants for
generously contributing their time. We appreciate the work of study staff
(Evan Beaupré, Niki Blakely, Jill Dalton, Vartouhi Jazmaji, Martin Kang, Kevin McCurley,
Andrew Thomas), hospital personnel (Barb Boychuk, Jan Buchanan, Doug Chisholm,
Nada Elfeki, Kishore Mulpuri) and our collaborators from the city (Peter Stary,
David Tomlinson, Barbara Wentworth) and community (Bonnie Fenton, David Hay,
Nancy Smith Lea, Fred Sztabinski).
RISK PERCEPTIONS FOR CYCLING
Figure 1.
Perceived risk by route type for 1,380 control sites in the Bicyclists’ Injuries and the Cycling Environment study
Perceived risk
N
Mean*
95% CI
Major street with shared lane & no parked cars
19
-0.21
(-0.54 - 0.11)
Major street
street, with
ith parked cars
265
-0.07
0 07
( 0 14 - 0.00)
(-0.14
0 00)
Major street, with no parked cars
232
0
(-0.08 - 0.08)
Major street with shared lane & parked cars
11
0.09
(-0.24 – 0.42)
Sidewalk
82
0.1
(-0.04 – 0.25)
Route type
Response frequency: How safe is this site?
Cycle
y track
19
0.18
(-0.15 – 0.52)
Major street with bike lane & parked cars
54
0.23
(0.06 – 0.41)
Residenal street designated bike route
100
0.25
(0.14 – 0.36)
Major street with bike lane & no parked cars
89
0.26
(0.13 – 0.39)
Mul-use paths, paved
109
0.36
(0.23 – 0.48)
Residenal street designated bike route, with traffic calming
110
0.41
(0.32 – 0.51)
Bike-only path
46
0.42
(0.25 – 0.60)
Residenal street
222
0.44
(0.37 – 0.51)
Mul-use paths, unpaved
22
0.66
(0.43 – 0.89)
0%
Very dangerous
20%
Somewhat dangerous
40%
60%
Neither safe nor dangerous
80%
100%
Somewhat safe
Very safe
* Response categories scaled from 1: very safe to -1: very dangerous for mean and CI calculaons
routes, bike lanes, bike paths, cycle tracks at roundabouts) have
been shown to reduce the risk of crashes and injuries compared
with cycling on road with traffic or off road with pedestrians.12
Most studies have used broad, route-type categories (e.g., on road,
off road, bike lane), although urban centres have many variations
on these themes. The Bicyclists’ Injuries and the Cycling Environment (BICE) study was the first to examine a detailed list of route
types that exist in North America.12 This evidence can suggest the
types of route that should be built to reduce risk.
However, decisions to cycle may be guided more by perceptions
than empirical data.13 Perceptions of risk or the likelihood that an
individual will experience a danger are influenced by both the
probability of an adverse event (e.g., the risk of a crash) and the
magnitude of the consequences (e.g., the severity of the injury).14,15
Risk perceptions vary by individual characteristics (sex, age, attitude) but are also heavily influenced by social and cultural conditions and interactions, and by the specific hazard.16 Perceived
reductions in risk may have greater than proportional effects on
encouraging or discouraging cycling, so it is especially important
that perceptions be taken into account.17 If there is discordance
between what is safe according to the empirical evidence and what
is perceived as safe, then even if protective infrastructure were to be
built, people might choose not to cycle. The objective of this paper
is to compare the perceived and observed injury risk of specific
route types, using data from the BICE study.
METHODS
Recruitment
The BICE study is a case-crossover study that recruited 690 adults
who visited emergency departments within 24 hours of incurring
an injury while cycling in Toronto or Vancouver. The study proto-
col and details have been published elsewhere.12,18 In brief, the
study population included injured cyclists aged 19 and over who
attended the emergency department of one of the study hospitals
in Vancouver (St. Paul’s, Vancouver General) or Toronto
(St. Michael’s, Toronto General, Toronto Western) between May 18,
2008, and November 30, 2009. Research staff at each hospital identified injured cyclists and provided contact information to the
study coordinators in their respective city (Vancouver or Toronto).
Introductory letters were sent to all potential participants, followed
by a telephone call from the study coordinator to invite participation and screen for eligibility. Up to 10 contact attempts were made
over a period of three months after their injuries. Cyclists were
excluded if they were injured outside of Toronto or Vancouver; did
not reside within the study cities; were unable to participate in an
interview (fatally injured, injured too seriously to communicate,
could not speak English or unable to remember the injury trip);
were injured while trick riding, racing, mountain biking, participating in a Critical Mass ride or on private property at the time of
the crash; were injured while riding a motorized bike (e.g., electric
or pedal-assisted bicycle), unicycle or tandem bike; or were already
enrolled in the study as a result of a previous injury. The study was
approved by research ethics boards at the universities of British
Columbia and Toronto, and at each of the five participating hospitals.
Interviews and site observations
Eligible participants were interviewed in person to identify personal
and trip characteristics. During the interview the injury location
and the injury trip route were recorded on a paper map. The injury
trip route distance was calculated using a digital map wheel. For
each injury site, two control sites (where no injury occurred) were
selected by multiplying a randomly generated proportion by the
Figure 2.
Observed relative risk (unadjusted odds ratio) versus perceived risk (mean score) by route type for 1,380 control sites in
the Bicyclists’ Injuries and the Cycling Environment study
High
Major street
with parked cars
(RR = 1.0)
Major street
with shared lane
& parked cars
Observed relaave risk
Mul-use paths,
paved
Sidewalk
Residenal street
designated bike route,
with traffic calming
Mul-use paths,
unpaved
Bike- only
path
Residenal street
designated bike route
Major street with bike lane
& parked cars
Major street with bike lane
& no parked cars
Residenal street
Low
Major street
& no parked cars
Major street
with shared lane
& no parked cars
Cycle track
(RR = 0.1)
Low
(Mean score = 0.7)
High
Perceived risk
(Mean score = - 0.2)
Legend
= high observed and perceived risks
= low observed and perceived risks
= higher observed than perceived risk
= lower observed than perceived risk
trip distance, then tracing the resulting distance along the route
using the map wheel. One site was selected to match the injury site
on the basis of location type, adjusted forward or backward, respectively, to match intersection or non-intersection location for control site, and the other was selected purely at random (i.e., not
deliberately matched to intersection status). Trained observers
blinded to the site status (injury or control) visited each site in the
field to collect data on site characteristics related to infrastructure
(the route type, intersecting streets, the presence of cycling infrastructure, intersection geometry), the physical environment (topography, visibility, street lighting, construction or other obstacles)
and usage (user volumes).
Measures and analysis
According to the street characteristics from site observations, we
classified route types into 14 categories, defined with input from
city bicycle transportation engineers and bicycling advocates.
Observed relative risk for this study was calculated using conditional logistic regression to estimate odds ratios (ORs) as a measure
of the relative risk of injury associated with different route types,
and these results have been reported elsewhere.12 In the present
paper, we used the unadjusted ORs based on a bivariate model of
the route type classification and site (1= injury site, 0=control site),
thereby not introducing other physical environment and usage
characteristics to the model. Perceived risk scores were based only
on the non-injury control sites, reducing bias by the injury event.
Perceived risk was calculated as the mean response of study participants to the question “How safe do you think this site was for
cyclists on that trip?” with a 5-point response scale: very safe, somewhat safe, neither safe nor dangerous, somewhat dangerous, very
dangerous. Responses were scaled from +1 (very safe) to 0 (neutral)
to -1 (very dangerous) for calculating and comparing mean scores
and 95% confidence intervals (CIs). We compared perceived risk
with observed ORs estimates using Pearson’s weighted correlation.
RESULTS
Study population
This study recruited 690 injured cyclists (414 in Vancouver, 276 in
Toronto), with a response rate of 93.1% of those confirmed to be
eligible or 66.5 % of those estimated as eligible. Details of the
recruitment process are reported in other manuscripts from the
study.12,18 Participants were predominately male (59.4%), young
(mean age=28), educated (75% with a post-secondary diploma or
degree) and employed full time (69.4%). Three quarters of the participants considered themselves experienced cyclists (530/690,
76.8%), 36.5% reporting that they cycled at least once a week in
the winter and 89.4% that they did so in the summer.
ceived and observed relative risk, as highlighted by the outliers in
Figure 2. The safest route type, cycle tracks, was perceived as higher risk than other bicycle-specific route types. Conversely, unpaved
multi-use paths were perceived as the safest route type, when in
fact they were observed to have a relative risk nearly as high as
major streets with no bike infrastructure or with shared lanes. Paved
multi-use paths were also more dangerous than perceived.
Study sites
We compared empirical data on observed relative risk with the perceptions of safety of different types of road infrastructure based on
the opinions, injury locations and travel patterns of injured cyclists
recruited from hospital emergency departments in Toronto and
Vancouver. It is known that route choices and decisions to cycle
are affected by perceptions of safety,1,13,19 and we found that perceptions largely aligned with observed evidence, with some exceptions. The injured cyclists in this study were weighted towards
experienced cyclists, and this may have allowed them to gauge the
risk of cycling infrastructure relatively well.
Perceptions of safety were aligned with published work on
cyclists’ route preferences. Previous work has indicated that people
have a preference for separated routes.20 The risk perceptions of this
population reflected these preferences. Separated route types were
perceived as safe, with unpaved multi-use paths, bike paths and
paved multi-use paths all rated very safe, and cycle tracks and sidewalks rated more neutrally. Again, corroborating published preferences, residential streets were perceived as safer than major streets.
Among major street types, those with bike lanes were perceived as
safer, as were streets with no car parking. The absence of parked
cars removes the risk of a cyclist colliding with an opening car door
and the need to deal with cars moving in and out of parking spaces.
The evidence on observed relative risk of certain separated route
types was not as closely aligned with risk perceptions as preferences
were. We found that cycle tracks, which separate cyclists and motor
vehicles on major city streets, carry about one-tenth the risk of typical major streets, yet these were perceived to have moderate risk.
It should be noted that there were only 19 sites of this type in the
control site sample, so the small sample leads to more uncertainty
in the OR estimate. At the time of this study, there were very few
cycle tracks in Vancouver and none in Toronto. Cycle tracks are
commonplace in European countries such as Denmark and the
Netherlands, where cycling is common and safety risks are low.21 In
Canada, cyclists are less familiar with this infrastructure so may
have had more trouble gauging its risk. Multi-use paths, while preferred20 and perceived as some of the safest types of route (as found
in this paper), offered only about a 25%-40% risk reduction compared with major streets. Given that reported safety concerns are
primarily around motor vehicles,1 a possibility may be that people
do not recognize the risk of injury from crashes with infrastructure,
cyclists, pedestrians or animals, or from falls due to slippery or
uneven surfaces along unpaved routes.6,7
As suggested in the risk literature, perceptions may be tied to the
severity of the consequences,14,16 that is, there may be a perceived
difference between a “risk of any injury” and a “risk of severe
injury”. Our injury study did not differentiate severity of injuries,
although all required attendance at an emergency department.
However, others have shown that most severe injuries and fatalities
do arise from collisions with motor vehicles.22,23 This potential for
Data from 1,380 control sites (two for each of the 690 injured
cyclists) were used in this analysis. Figure 1 lists the 14 route categories and their frequencies. Ubiquitous route types were common
in the dataset (major streets with no bike infrastructure and residential streets), whereas certain specialized bicycle infrastructure
was relatively rare (major streets with shared lanes and cycle tracks,
i.e., bike lanes alongside major streets but separated by a physical
barrier).
Perceived risk
Figure 1 also summarizes perceived risk responses by mean scores.
The four route types with the highest perceived risks were major
streets (with and without parking) and either shared lanes (with
cars, buses or high occupancy vehicles) or no bicycle infrastructure.
The mean score for sidewalks was 0.10 (neither safe nor dangerous), although the response frequency indicates that sites were split
between being perceived as safe and being perceived as dangerous.
The following route types were considered safe at more than 50%
of sites of that type: off-street multi-use paths; residential streets;
off-street bike paths; major streets with bike lanes; and cycle tracks.
Observed risk
The results of observed relative risk are reported in detail in another manuscript from this study.13 To briefly summarize those results,
the reference category was major street with parked cars and no
cycling infrastructure (OR=1). All other route types had lower risk
of injury. Cycle tracks had the lowest relative risk, at almost one
tenth of the risk (OR=0.12). Route types with about half the relative
risk (ORs from 0.44 to 0.59) were residential streets (designated as
bike routes or not, with or without traffic calming), major streets
with bike lanes (with or without parking) and off-street paths for
bikes only. Multi-use paths, sidewalks and other major street configurations (with shared lanes or with no bike infrastructure) all
had higher relative risk (ORs from 0.63 to 0.78).
Comparing perceived and observed relative risk
A comparison of perceived risk and observed relative risk is presented in Figure 2. Four of the more dangerous route types were
also perceived as unsafe: major streets with shared lanes or without
any bicycle infrastructure. Many safer route types were also perceived as safer: residential streets (designated as bike routes or not,
with or without traffic calming), major streets with bike lanes (with
or without parking) and off-street paths for bikes only. The rank
ordering of risk was not the same for perceived and observed relative risk (Pearson’s weighted correlation=-0.68, p=0.007, negative
because high perceived risk had negative safety scores, and high
observed relative risk had the highest ORs). This moderate correlation was influenced by some major discrepancies between per-
DISCUSSION
more severe injuries may have been considered in cyclists’ higher
risk perception for cycle tracks alongside major streets and lower
risk perceptions for multi-use paths away from traffic. The importance of being away from traffic to reduce perceptions of risk has
been shown by others,24 and this tendency to favour quiet routes
may explain the discrepancies between observed and perceived risk
ratings for both cycle tracks and multi-use paths. However, risk perceptions in this study were not unrefined, as sidewalks were appropriately perceived as intermediate risk. This parallels the empirical
risks from our injury study and other literature demonstrating that
riding on sidewalks is more dangerous than on bicycle-specific
infrastructure and many road types.25
The BICE study was a case-crossover study, a robust methodology
that compares an injury site with control site(s) for the same individual and trip. Use of this methodology enabled an inquiry into the
independent effects of route infrastructure while accounting for challenges facing injury research around exposure to risk and the confounding effects of unmeasured individual and trip characteristics.18
This provided rigorous estimates for observed relative risk for 14 different route types. While this study included the most diversity of
infrastructure in any cycling study we are aware of, it was limited to
the existing types of infrastructure in Vancouver and Toronto.
We used only data on the BICE control sites for the analysis of
perceived risk to reduce bias introduced by the injury event. When
we analyzed the reports of perceived safety at injury sites, we found
that the mean reported risk at injury sites was typically 0.3-0.4 units
greater than reported for control sites of the same type. This could
be bias resulting from the fact that the participants had experienced
an injury there, or indeed it could be that conditions at these sites
in particular were more dangerous than was the average site of that
type. Given this unknown, these were not included in analyses. It
may also be that injured individuals have heightened perceptions
of risk at any site (injury or control) following an injury event, as
compared with the response they may have made in the absence of
an injury. Our analysis focused on the order of perceived risk, more
so than on absolute scores.
Additionally, this study considered risk of injury from crashes,
not safety risks associated with personal crime or bicycle theft, or
health risks from exposure to air pollution. Whether study participants restricted the interpretation of our safety question to injury
risk is not known, although it seems likely that a cyclist who was
recently injured, as in this study, would have that aspect of safety
at the forefront of their thoughts. As the interview question referred
to “this site”, participants may have reflected on other factors (traffic speed and volume, vehicle type, weather) beyond simply the
road infrastructure, though some of these features are also strongly related to route type (e.g., traffic volume on major streets was
about 20 times higher than on residential streets).
This study included injuries severe enough to require an emergency department visit. Minor injuries were not captured, nor were
fatal injuries or people with severe head injuries who could not
recall the study trip (although these represented only 1.6% of those
known to be eligible, including two fatalities during the study). A
future analysis of the BICE study data will examine risk factors for
more severe injury, in which we will query the role of route infrastructure and the involvement of motor vehicles.
Contributions
Our results have a number of implications for practice. The data
on perceived and observed relative risk can guide municipalities on
the types of infrastructure that are safe according to both criteria
and are therefore likely to both attract cyclists and keep them safe.
The findings on discrepancies between perceived safety and
observed relative risk carry different implications for practice and
advocacy. This highlights the need to inform public opinion on
route safety, perhaps through education programs and social media,
in order to encourage cycling and use of the safest possible infrastructure. Improving cyclists’ knowledge about the comparative risk
of infrastructure types may reduce injury incidence by influencing
their route choices or their risk-taking behaviour while using more
dangerous types of cycling infrastructure.
REFERENCES
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
Winters M, Davidson G, Kao DN, Teschke K. Motivators and deterrents of
bicycling: Comparing influences on decisions to ride. Transportation
2011;38(1):153-68.
Bhat CR, Sener IN, Eluru N. Who are bicyclists? Why and how much are they
bicycling? Transp Res Record 2009;2134:63-72.
Beck LF, Dellinger AM, O’Neil ME. Motor vehicle crash injury rates by mode
of travel, United States: Using exposure-based methods to quantify differences. Am J Epidemiol 2007;166(2):212-18.
Transport Canada. Canadian Motor Vehicle Traffic Collision Statistics: 2006.
TP 3322. Available at: http://www.tc.gc.ca/eng/roadsafety/tp-tp3322-2006menu-586.htm (Accessed November 12, 2011).
Statistics Canada. Commuting Patterns and Places of Work of Canadians,
2006 Census. Catalogue no. 97-561-X, 2006 Available at: http://www12.statcan.ca/census-recensement/2006/as-sa/97-561/tables-tableaux-notes-eng.cfm
(Accessed May 1, 2011).
Babul S, Frendo T, Winters M, Brubacher J, Chipman M, Chisholm D, et al.
Injuries to adult cyclists in Toronto and Vancouver: Describing the circumstances as a first step towards injury prevention. Safety 2010 World Conference, London, England, 2010.
de Geus B, Vandenbulcke G, Int Panis L, Thomas I, Degraeuwe B, Cumps E,
et al. A prospective cohort study on minor accidents involving commuter
cyclists in Belgium. Accid Anal Prev 2012;45:683-93.
Joshi MS, Smith GP. Cyclists under threat: A survey of Oxford cyclists’ perceptions of risk. Health Educ J 1992;51(4):188-91.
Stutts JC, Williamson JE, Whitley T, Sheldon FC. Bicycle accidents and
injuries: A pilot study comparing hospital- and police-reported data.
Accid Anal Prev 1990;22(1):67-78.
Juhra C, Wieskötter B, Chu K, Trost L, Weiss U, Messerschmidt M, et al.
Bicycle accidents – Do we only see the tip of the iceberg? A prospective
multi-centre study in a large German city combining medical and police data.
Injury November 18, 2011; Epub ahead of print.
Pucher J, Dill J, Handy S. Infrastructure, programs, and policies to increase
bicycling: An international review. Prev Med 2010;50:S106-S125.
Teschke K, Harris M, Reynolds C, Winters M, Babul S, Chipman M, et al. Route
infrastructure and the risk of injuries to bicyclists – a case-crossover study.
Am J Public Health (in press).
Dill J, Voros K. Factors affecting bicycling demand: Initial survey findings
from the Portland, Oregon, region. Transp Res Record 2007;2031:9-17.
Leiss W, Chociolko C. Risk and Responsibility. Montreal & Kingston: McGillQueen’s University Press, 1994.
Slovic P. Perception of risk. Science 1987;236(4799):280-85.
Sjöberg L, Moen B, Rundmo T. Explaining Risk Perception. An Evaluation of the
Psychometric Paradigm in Risk Perception Research. Trondheim, Norway:
Rotunde publikasjoner, 2004.
Noland R. Perceived risk and modal choice: Risk compensation in transportation systems. Accid Anal Prev 1995;27(4):503-21.
Harris MA, Reynolds CC, Winters M, Chipman M, Cripton PA, Cusimano
MD, et al. The Bicyclists’ Injuries and the Cycling Environment study: A protocol to tackle methodological issues facing studies of bicycling safety.
Inj Prev 2011;17(5):e6.
Dill J. Bicycling for transportation and health: The role of infrastructure.
J Public Health Policy 2009;30:S95-S110.
Winters M, Teschke K. Route preferences among adults in the near market
for bicycling: Findings of the Cycling in Cities Study. Am J Health Promot
2010;25(1):40-47.
Pucher J, Buehler R. Making cycling irresistible: Lessons from the Netherlands,
Denmark and Germany. Transport Rev 2008;28(4):495-528.
Rivara FP, Thompson DC, Thompson RS. Epidemiology of bicycle injuries and
risk factors for serious injury. Inj Prev 1997;3(2):110-14.
23. Morgan AS, Dale HB, Lee WE, Edwards PJ. Deaths of cyclists in London:
Trends from 1992 to 2006. BMC Public Health 2010;10:699.
24. Parkin J, Wardman M, Page M. Models of perceived cycling risk and route
acceptability. Accid Anal Prev 2007;39(2):364-71.
25. Reynolds CCO, Harris MA, Teschke K, Cripton PA, Winters M. The impact of
transportation infrastructure on bicycling injuries and crashes: A review of
the literature. Environ Health Perspect 2009;8:47.
RÉSUMÉ
Objectif : Les préoccupations quant à la sécurité ont un effet dissuasif
sur le cyclisme. L’étude BICE (Bicyclists’ Injuries and the Cycling
Environment) a quantifié le risque de blessure associé à 14 types de routes
– des sentiers hors route aux grandes artères. Lorsqu’il s’agit du risque de
blessure, il peut y avoir discordance entre les preuves empiriques et les
perceptions. Quand c’est le cas, même si l’on construit des infrastructures
de protection, les gens peuvent ne pas se sentir suffisamment en sécurité
pour faire du vélo. Notre article porte sur la relation entre le risque de
blessure subjectif et observé.
Méthode : L’étude BICE est une étude de type « case-crossover » pour
laquelle nous avons recruté 690 cyclistes adultes s’étant rendus aux
services d’urgence de Toronto et de Vancouver après un accident de vélo.
Nous avons calculé le risque observé en comparant le type de route sur le
lieu de l’accident avec les types de routes sur des lieux sélectionnés au
hasard le long du même parcours. Le risque subjectif était la réponse
moyenne des participants de l’étude à la question : « Quel était le niveau
de sécurité de cet endroit pour les cyclistes durant ce trajet? »; les
réponses ont été classées de +1 (très sûr) à -1 (très dangereux). Les scores
de risque subjectif n’ont été calculés que pour les lieux témoins (sans
accident) afin de réduire le biais induit par l’accident.
Résultats : Les types de routes présentant le plus grand risque subjectif
étaient les grandes artères avec voie partagée, sans voitures garées (score
moyen = -0,21, intervalle de confiance [IC] de 95 % : -0,54-0,11), suivies
des grandes artères sans infrastructures cyclables (-0,07, IC -0,14-0,00).
Les routes perçues comme étant les plus sûres étaient les sentiers multiusages asphaltés (0,66, IC 0,43-0,89), les rues résidentielles (0,44,
IC 0,37-0,51), les pistes cyclables (0,42, IC 0,25-0,60) et les rues
résidentielles marquées pour les bicyclettes et comportant des mesures
de modération de la circulation (0,41, IC 0,32-0,51). La plupart des types
de routes perçues comme étant plus dangereuses étaient de fait plus
dangereuses dans notre étude; de même, la plupart des types de routes
perçues comme étant moins dangereuses l’étaient effectivement. Des
divergences ont été notées pour les pistes cyclables (le risque subjectif
étant plus élevé que le risque observé) et pour les sentiers multi-usages
(le risque observé étant plus élevé que le risque subjectif).
Conclusions : Le choix d’une route et la décision de faire du vélo sont
influencés par les perceptions de la sécurité, et nous avons constaté que
ces perceptions correspondent généralement à la sécurité objective.
Toutefois, les perceptions de certains types de voies séparées concordent
moins avec la réalité. Des programmes de sensibilisation du public et
dans les médias sociaux pourraient faire en sorte que les perceptions de la
sécurité des routes par le public reflètent les données probantes.
Mots clés : sécurité; transports; traumatismes; conception de
l’environnement
Associations Between Children’s Diets and Features of Their
Residential and School Neighbourhood Food Environments
Andraea Van Hulst, RN, MSc,1,2 Tracie A. Barnett, PhD,1,2 Lise Gauvin, PhD,1,3,4 Mark Daniel, PhD,1,5,6
Yan Kestens, PhD,1,3 Madeleine Bird, BA,1,2 Katherine Gray-Donald, PhD,7 Marie Lambert, MD2,8
ABSTRACT
Objectives: Among studies of the built environment, few examine neighbourhood food environments in relation to children’s diets. We examined the
associations of residential and school neighbourhood access to different types of food establishments with children’s diets.
Methods: Data from QUALITY (Quebec Adipose and Lifestyle Investigation in Youth), an ongoing study on the natural history of obesity in 630 Quebec
youth aged 8-10 years with a parental history of obesity, were analyzed (n=512). Three 24-hour diet recalls were used to assess dietary intake of
vegetables and fruit, and sugar-sweetened beverages. Questionnaires were used to determine the frequency of eating/snacking out and consumption of
delivered/take-out foods. We characterized residential and school neighbourhood food environments by means of a Geographic Information System.
Variables included distance to the nearest supermarket, fast-food restaurant and convenience store, and densities of each food establishment type
computed for 1 km network buffers around each child’s residence and school. Retail Food Environment indices were also computed. Multivariable
logistic regressions (residential access) and generalized estimating equations (school access) were used for analysis.
Results: Residential and school neighbourhood access to supermarkets was not associated with children’s diets. Residing in neighbourhoods with lower
access to fast-food restaurants and convenience stores was associated with a lower likelihood of eating and snacking out. Children attending schools in
neighbourhoods with a higher number of unhealthful relative to healthful food establishments scored most poorly on dietary outcomes.
Conclusions: Further investigations are needed to inform policies aimed at shaping neighbourhood-level food purchasing opportunities, particularly for
access to fast-food restaurants and convenience stores.
Key words: Built environment; children; diet; food environment; residential neighbourhood; school neighbourhood; QUALITY cohort
I
n recent years, the role of neighbourhoods has been increasingly investigated with respect to obesity in children.1-3 Neighbourhood built environments may promote childhood obesity by
favouring antecedent behaviours, including physical inactivity and
unhealthful diets. Compared with physical activity, fewer studies
have addressed children’s diets.1
Most studies examining associations between local neighbourhood availability of food establishments and residents’ diets have
focused on adults.4 Overall, findings from studies involving children are less consistent, notably for associations between access to
supermarkets and vegetable and fruit (V&F) intake.5-7 Greater access
to convenience stores, which typically offer limited fresh produce,
has been found to be associated with lower V&F intake5,7 and higher intake of sweet/salty snacks6 and sugar-sweetened beverages8 in
youth. Although some studies have reported associations between
the availability of fast-food restaurants near children’s residence
and their diets,7,8 others do not support such findings.6,9,10 Given
the conflicting results in the literature, there is a need to clarify the
relation between neighbourhood food environments and children’s
diets.
In addition to residential neighbourhoods, school neighbourhood environments are relevant activity spaces and should be
investigated in relation to obesity-related behaviours in children.11,12 During the academic year, travel to and from school
exposes children to school neighbourhood food environments.
Recently, policies have targeted in-school food environments, but
initiatives aimed at regulating food opportunities in school neighbourhoods have yet to be widely implemented. Fast-food restaurants and convenience stores are known to cluster within short
distances from schools.13,14 However, it is not clear to what extent
the availability of the latter is associated with children’s diet.2,9
Author Affiliations
1.
2.
3.
4.
Département de médecine sociale et préventive, Université de Montréal, Montréal, QC
Centre de recherche du Centre Hospitalier Universitaire Sainte-Justine, Montréal, QC
Centre de Recherche du Centre Hospitalier de l’Université de Montréal, Montréal, QC
Centre de Recherche Léa-Roback sur les Inégalités Sociales de Santé de Montréal,
Université de Montréal, Montréal, QC
5. School of Health Sciences, University of South Australia, Adelaide, Australia
6. Department of Medicine, St Vincent’s Hospital, The University of Melbourne,
Melbourne, Australia
7. School of Dietetics and Human Nutrition, McGill University, Montréal, QC
8. Département de pédiatrie, Université de Montréal, Montréal, QC
Correspondence: Andraea Van Hulst, Centre de recherche du CHU Sainte-Justine,
5757 Ave. Decelles, suite 100, Montréal, QC H3S 2C3, Tel: 514-345-4931 (ext. 3271),
Fax: 514-345-4801, E-mail: [email protected]
Acknowledgements: The QUALITY study is funded by the Canadian Institutes of
Health Research (CIHR), the Heart and Stroke Foundation of Canada (HSFC) and the
Fonds de la recherche en santé du Québec. The QUALITY Residential and School Built
Environment complementary studies were funded by the HFSC and the CIHR,
respectively. A.Van Hulst received support from a CIHR/HSFC Training Grant in
Population Intervention for Chronic Disease Prevention, and a doctoral scholarship
from the Fonds de la recherche en santé du Québec; T. Barnett and Y. Kestens are
research scholars with Fonds de la recherche en santé du Québec; L. Gauvin holds a
CIHR/Centre de Recherche en Prévention de l’Obésité Chair in Applied Public Health
on Neighbourhoods, Lifestyle, and Healthy Body Weight; and M. Bird received a
Masters’ scholarship from the Fondation du CHU Sainte-Justine. Dr. Marie Lambert
passed away on February 20, 2012. Her leadership and devotion to the QUALITY
cohort will always be remembered and appreciated.
NEIGHBOURHOOD FOOD ENVIRONMENTS AND DIET
The aim of this study was to determine whether features of residential and school neighbourhood food environments were associated with children’s dietary intake (V&F and sugar-sweetened
beverages) and selected dietary behaviours (eating/snacking out
and consuming delivered/take-out food).
METHODS
Participants were drawn from the QUALITY (Quebec Adipose and
Lifestyle Investigation in Youth) study, an ongoing longitudinal
investigation of the natural history of obesity and cardiovascular
risk in youth with a history of parental obesity. Recruitment flyers
were distributed to parents of children in Grades 2 to 5 in 1,040
primary schools (89% of schools approached) located within 75 km
of each of Montreal, Quebec City and Sherbrooke, QC. Of 3,350
interested families who contacted the research coordinator, 1,320
met the study inclusion criteria. Eligibility criteria required participating children to be Caucasian, aged 8-10 years at recruitment
and to have at least one obese biological parent (i.e., body mass
index [BMI] ≥30 kg/m2 and/or waist circumference >102 cm in men
and >88 cm in women, based on self-reported measurements of
height, weight and waist circumference) and both biological parents available to participate at baseline. Of eligible families, a total
of 630 (48% of eligible families composed of the participating child
and two biological parents) completed the baseline visit between
September 2005 and December 2008. Baseline data collection
involved a clinic visit during which questionnaires were completed and biological and physiological measurements obtained, as well
as follow-up telephone interviews. Written informed consent was
obtained from parents, and assent was provided by children. The
ethics review boards of Centre Hospitalier Universitaire SainteJustine and Laval University approved the study. A detailed description of the study design and methods is available elsewhere.15
Characteristics of the built and social environments in children’s
residential neighbourhood were obtained for the study baseline
using a Geographic Information System (GIS) for 512 children
residing in the Montreal Census Metropolitan Area (CMA). Of
these, 506 attended some 296 schools located within the Montreal
CMA, for which school neighbourhood GIS data were also
obtained.
Dietary assessment
Children’s dietary intake was measured using mean values of three
24-hour diet recalls conducted by trained dieticians on nonconsecutive days, including one weekend day.16 Data from recalls
were available for 498 participants considered in this study. Except
in unusual circumstances, the recalls were collected within a 4-week
period after the baseline clinic visit. Diet recall interviews were done
by telephone with the child and then confirmed with the parent
who prepared the meals.
Foods reported on the recalls were entered into CANDAT
(London, ON) and converted to nutrients using the 2007b Canadian
Nutrient File.17 Daily servings of V&F were based on portion sizes
from Canada’s Food Guide and include V&F juices. A dichotomous
variable was developed on the basis of recommended servings of
V&F for children aged 8-10 years: ≥5 servings/day vs. less.18 Intake
of sugar-sweetened beverages was computed as the mean daily
number of millilitres of soft drinks and other sugar-sweetened
drinks, but excluding juices made from real fruits. Given a sub-
stantial positive skewness in its distribution, sugar-sweetened beverage intake was dichotomized to >50 mL/day (approximately one
can of soft drink per week) vs. less.
Two additional measures of children’s diets were obtained from
a questionnaire administered to the child during the clinic visit:
having a meal or snack in a food establishment at least once in the
previous week and consuming delivered or “take-out” food at least
once in the previous week.
Neighbourhood assessment
The exact addresses of each participating child’s residence and
school were geocoded. The availability of food establishments within the residential and school neighbourhood environment was
measured using a GIS, which included data from an exhaustive list,
acquired from Tamec Inc., of businesses and services located in the
region in May 2005. The business name, address, postal code and
Standard Industry Classification code were available. A validity
study of food establishments from this list, verified by onsite field
visits, showed good agreement (0.77), sensitivity (0.84) and positive
predictive value (0.90).19 All businesses were geocoded using
GeoPinPoint™, version 2007.3 (DMTI Spatial Inc.). Types of food
establishment included in this study were supermarkets, fast-food
restaurants,14 convenience stores and specialty food stores (e.g., bakeries, fruit and vegetables, gourmet, meat and fish markets).
Neighbourhood food environments were described by proximityand density-based indicators. Proximity measures were established
using ArcGIS Network Analyst (Esri, Redlands, CA) and defined as
the road-network distance between the child’s residence and the
nearest supermarket, fast-food restaurant and convenience store,
and between the child’s school and the nearest of each food establishment type. Because of highly skewed distributions, indicators
were categorized into tertiles corresponding to farthest, intermediate and shortest distances. Kernel density was used to estimate the
average density of each type of food establishment within 1 km
street network buffers centred on 1) the child’s residence and 2) the
child’s school. Kernel density estimations are frequently used in
geography to evaluate the local density of point-based data20 and
have been used previously to describe neighbourhood access to
food establishments.21 A quartic kernel function was used with
adaptive bandwidth composed of 1% of the observations for each
type of food establishment (n=1,929 for convenience stores,
n=1,118 for fast-food restaurants and n=828 for supermarkets) and
cell spacing of 100 m. Exposure categories for each type of food
establishment were based on tertiles corresponding to lowest, intermediate and highest densities.
Additionally, a Retail Food Environment Index (RFEI) was computed.22 This index is based on the ratio of the number of fast-food
restaurants and convenience stores to supermarkets and specialty
food stores. Higher scores are indicative of neighbourhoods characterized by a higher number of unhealthful relative to healthful
options. The RFEI was computed for 1 km network buffers and for
3 km radius circular buffers centred on each of the residential and
school locations. A larger buffer was examined to capture greater
variation among neighbourhoods in RFE indices. The index
was subsequently categorized according to the approximate
75th percentile of each variable’s distribution, corresponding to
cut-offs of ≥2.0 vs. less for 1 km buffers and ≥2.5 vs. less for 3 km
buffers.
Table 1.
Characteristics of Participants Residing Within the
Montreal CMA (n=512) From the Quebec Adipose
and Lifestyle Investigation in Youth (QUALITY) Study
Characteristic
Mean (SD) age, years
9.6 (0.9)
% (n) of male sex
54.5 (279)
Mean (SD) annual household income, $*
43,063 (18,722)
Highest level of education of either parent, % (n)
2 parents with secondary school or less
8.3 (42)
1 or 2 parents with technical/vocational/trade degree
38.5 (196)
1 or 2 parents with university degree
53.2 (271)
29.5 (6.6)
Mean (SD) BMI of mothers, kg/m2
2
30.8 (5.6)
Mean (SD) BMI of fathers, kg/m
Mean (SD) no. of daily servings of V&F
4.3 (2.1)
≥5 servings of V&F per day, % (n)
33.7 (168)
>50 mL of sugar-sweetened beverages per day, % (n)
58.0 (289)
Eat/snack out at least once per week, % (n)
43.8 (224)
Delivered/take-out food at least once per week, % (n)
35.0 (179)
Residential neighbourhood
2
2715 (1926-3815)
Median (IQR) population density per km
% aged ≥15 years with no high school diploma, mean (SD) 32.6 (9.0)
% aged ≥15 years who are employed, mean (SD)
67.0 (8.3)
Mean (SD) total income of households, $
85,793 (23,197)
Median (IQR) walking distance from residence to
school, metres
1121 (631-2535)
Proximity measures (distance to nearest), metres
Supermarket, median (IQR)
1375 (739-2434)
Fast-food restaurant, median (IQR)
1326 (784-2256)
Convenience store, median (IQR)
779 (425-1327)
2
Kernel density measures (for 1 km network buffer), no./km
0.08 (0.03-0.2)
0.2 (0.08-0.8)
0.3 (0.1-1.0)
Retail Food Environment Index
1 km network buffer, median (IQR)
1.0 (0-2.0)
3 km circular buffer, median (IQR)
1.8 (1.2-2.5)
School neighbourhood†
2990 (2093-4087)
Median (IQR) population density per km2
% aged ≥15 years with no high school diploma, mean (SD) 32.9 (9.1)
% aged ≥15 years who are employed, mean (SD)
64.0 (8.1)
Mean (SD) total income of households, $
81,478 (20,793)
Proximity measures (distance to nearest), metres
1008 (540-1999)
950 (572-1889)
541 (311-931)
2
Kernel density measures (for 1 km network buffer), no./km
0.1 (0.03-0.3)
Fast-food restaurant median (IQR)
0.3 (0.1-1.0)
Convenience store median (IQR)
0.5 (0.2-1.6)
Retail Food Environment Index
1 km network buffer, median (IQR)
0.8 (0-1.8)
3 km circular buffer, median (IQR)
1.7 (1.2-2.4)
* Adjusted for the number of people living in the household.
† School neighbourhood data available for 296 schools localized within the
Montreal CMA attended by 506 QUALITY study children (6 attended a
school outside the study area).
CMA=Census Metropolitan Area; BMI=body mass index; V&F=vegetables and
fruit; IQR=inter-quartile range.
Other neighbourhood-level measures included a material deprivation index computed from 2006 Census data.23 The index combines the proportion of people with no high school diploma, the
proportion who are employed and the average income, for people
aged ≥15 years in census dissemination areas. Population-weighted
proportions of dissemination areas overlapping 1 km street network buffers centred on resident’s location were computed. The
index was classified into quintiles of lowest to highest deprivation.
A material deprivation index for school neighbourhood was computed using the same approach. Population density for both residential and school neighbourhood environments was computed
from 2006 Census data for 1 km street network buffers. A median
split categorization was used for measures of population density.
Individual socio-demographic measures
Individual-level data used as adjustment variables included child’s
age and sex, and mother’s BMI. Highest parental educational attainS50 REVUE CANADIENNE DE SANTÉ PUBLIQUE • VOL. 103, SUPPLÉMENT 3
ment (2 parents with secondary school or less, ≥1 parent with technical/vocational/trade degree, ≥1 parent with university degree) and
total annual household income adjusted for the number of people
living in the household were obtained from parent-completed questionnaires during the clinic visit.
Analysis
This study was not designed to allow multilevel analyses of participants nested into neighbourhoods; instead, an ego-centred
approach was used whereby individual neighbourhood measures
were computed for each child’s residential and school locations.24
Moreover, no evidence of spatial autocorrelation resulting from the
dependency of properties within geographic spaces was found, indicating that nearby entities did not share more similarities than entities that were further apart (data not shown).
Unadjusted associations among indicators of residential neighbourhood food environment and dietary outcomes were examined
using logistic regression. Subsequently, multivariable associations
were analyzed adjusting for child’s age and sex, as well as for potential confounders, namely parental education, household income,
residential neighbourhood material deprivation and residential
population density (as a measure of level of urbanicity). For analyses involving school neighbourhoods, generalized estimating equations (GEE) with a logit link function and with an independent
working correlation structure were used to allow for clustering of
dietary outcomes among children attending the same schools. Multivariable GEE models were adjusted for child’s age, sex, parental
education, household income, school neighbourhood material deprivation and school neighbourhood population density. Given the
high correlations between proximity-based indicators and between
density-based indicators of each type of food establishment (r=0.7
to 0.9), each variable was examined in separate models for residential and school neighbourhoods using the “best access” (i.e.,
closest or densest tertile) as the reference category. For RFE indices,
values below the cut-offs were used as the reference category. Odds
ratios (OR) and 95% confidence intervals (CIs) are presented. All
analyses were conducted using SAS, version 9.2 (Cary, NC).
In secondary analyses, we restricted the sample to children who
lived >1.5 km from their school, i.e., those who were more likely to
have distinct residential and school neighbourhood food environments, since there would be minimal overlap between respective
1 km network buffers centred on each location. Associations
between the density of food establishments and dietary outcomes
were examined in this subgroup in an attempt explore which of
the residential or school neighbourhood food environment features
were most strongly associated with dietary outcomes.
RESULTS
Overall, 34% of the 512 children consumed the recommended
daily intake of ≥5 servings of V&F per day (average of 4.3 servings),
58% drank >50 mL of sugar-sweetened beverages daily, 44% had a
meal/snack in a food establishment, and 35% consumed delivered/take-out foods at least once per week (Table 1). Overall, supermarkets, fast-food restaurants and convenience stores were more
accessible around schools than around residences, as shown by
shorter distances to and higher densities of each type of food establishment in school neighbourhoods. Thirty-eight percent (n=193)
lived >1.5 km from their school.
Table 2.
Covariate-adjusted Associations (OR and 95% CI) Between Measures of the Residential Neighbourhood Food Environment
and Dietary Outcomes in the QUALITY Study*
≥5 Servings of
V&F/Day
>50 mL Sugarsweetened
Beverages/Day
(n=493)
Eating/Snacking
Out
≥Once/Week
(n=506)
Delivered/Takeout Food
≥Once/Week
(n=506)
1.09 (0.62-1.91)
1.07 (0.65-1.74)
1
0.82 (0.48-1.39)
0.84 (0.52-1.35)
1
1.04 (0.62-1.73)
1.12 (0.71-1.77)
1
0.96 (0.56-1.65)
1.47 (0.92-2.36)
1
1.39 (0.81-2.40)
1.27 (0.77-2.10)
1
0.82 (0.49-1.37)
0.98 (0.61-1.58)
1
1.03 (0.63-1.68)
1.08 (0.69-1.71)
1
1.03 (0.61-1.73)
1.40 (0.87-2.24)
1
0.99 (0.57-1.72)
0.98 (0.59-1.63)
1
0.85 (0.50-1.44)
0.87 (0.54-1.40)
1
1.15 (0.70-1.90)
1.23 (0.78-1.96)
1
0.93 (0.55-1.56)
1.02 (0.63-1.64)
1
1.11 (0.63-1.93)
0.87 (0.52-1.48)
1
1.20 (0.70-2.05)
1.38 (0.84-2.29)
1
0.63 (0.37-1.05)‡
0.78 (0.48-1.26)
1
0.91 (0.53-1.58)
1.40 (0.85-2.29)
1
1.22 (0.68-2.22)
1.01 (0.59-1.74)
1
1.19 (0.67-2.11)
1.24 (0.74-2.08)
1
0.52 (0.30-0.91) †
0.60 (0.36-0.99)†
1
1.11 (0.63-1.98)
1.10 (0.66-1.84)
1
1.02 (0.55-1.91)
1.17 (0.68-2.04)
1
1.25 (0.69-2.27)
1.19 (0.70-2.03)
1
0.44 (0.25-0.80)†
0.60 (0.36-1.02) ‡
1
0.93 (0.51-1.70)
1.15 (0.68-1.95)
1
0.90 (0.58-1.42)
1
0.93 (0.61-1.43)
1
1.01 (0.67-1.52)
1
1.35 (0.89-2.05)
1
0.77 (0.49-1.21)
1
0.94 (0.62-1.44)
1
0.88 (0.59-1.33)
1
1.22 (0.80-1.87)
1
(n=493)§
Proximity measures
Model 1 – distance to nearest supermarket
Farthest (>2000 m)
Intermediate (965 to 2000 m)
Shortest (<965 m)
Model 2 – distance to nearest fast-food restaurant
Farthest (>1835 m)
Shortest (<940 m)
Model 3 – distance to nearest convenience store
Farthest (>1090 m)
Shortest (<545 m)
Density measures
Model 4 – density of supermarkets
Lowest
Intermediate
Highest
Model 5 – density of fast-food restaurants
Lowest
Intermediate
Highest
Model 6 – density of convenience stores
Lowest
Intermediate
Highest
Retail food environment measures
Model 7 – 1 km buffer RFE Index
≥2 (27.3%)
<2 (72.7%)
≥2.5 (26.2%)
<2.5 (73.8%)
* Separate logistic regression models for each main exposure and each outcome, adjusted for child’s age, sex, parental education, household income, residential
neighbourhood material deprivation and residential population density.
§ When treated as a continuous outcome, farthest (vs. shortest) distance to the nearest fast-food restaurant was associated with V&F intake (Beta=0.50, 95% CI:
0, 1.00); and 3 km RFE index ≥2.5 (vs. less) was associated with V&F intake (Beta=-0.40, 95% CI: -0.81, 0.005).
† p≤0.05; ‡p≤0.10.
OR=odds ratio; CI=confidence interval; V&F=vegetables and fruit; RFE=retail food environment.
Tables 2 and 3 show covariate-adjusted associations of proximity, density and retail food environment measures with children’s
dietary outcomes for both residential and school neighbourhood
environments respectively. Living in a residential neighbourhood
with a lower density of fast-food restaurants was associated with a
48% (OR=0.52, 95% CI: 0.30-0.91) and 40% (OR=0.60, 95% CI:
0.36-0.99) lower likelihood of eating/snacking out, for lowest and
intermediate densities respectively. Similar associations were found
for convenience stores, the lowest density compared with the highest density indicating a 56% (OR=0.44, 95% CI: 0.25-0.80) lower
likelihood of eating/snacking out. Residential neighbourhood proximity-based indicators were not associated with children’s diets,
nor were residential RFE indices. Access to food establishments in
the school environment was only marginally associated with
dietary outcomes (Table 3). For example, intermediate (vs. shortest) distance between attended school and the nearest fast-food
restaurant was associated with an increased likelihood of consuming recommended servings of V&F (p=0.08). Similarly, attending
schools in neighbourhoods with the lowest density of supermarkets (vs. highest density) was associated with a decreased likelihood of eating/snacking out (p=0.08); an intermediate density of
supermarkets (vs. highest density) was associated with an increased
likelihood of consuming sugar-sweetened beverages (p=0.07); and
intermediate density of fast-food restaurants (vs. highest density)
was associated with an increased likelihood of consuming delivered/take-out foods (p=0.09).
The residential neighbourhood RFE indices were not associated
with dietary outcomes (Table 2). Attending a school in a neighbourhood with a 3 km buffer RFE Index ≥2.5 (i.e., 2.5 fast-food
restaurants/convenience stores for 1 supermarket/specialty store)
was associated with a 61% (OR=1.61, 95% CI: 1.01-2.56) greater
likelihood of consuming sugar-sweetened beverages, after adjustment for individual and neighbourhood covariates (Table 3). Similarly, an elevated RFEI within 1 and 3 km buffers around schools
was marginally associated with a lower likelihood of consuming
recommended servings of V&F.
Among children living >1.5 km from their school, lowest (vs.
highest) school neighbourhood density of fast-food restaurants was
associated with a higher likelihood of consuming recommended
servings of V&F, and intermediate (vs. highest) school neighbourhood density of fast-food restaurants was associated with a higher
likelihood of consuming delivered/take-out food (Table 4). The residential density of convenience stores remained positively associated with eating/snacking out.
Last, when V&F intake was treated as a continuous variable using
linear regression models, children living farthest from fast-food
restaurants had a 0.5 additional serving of V&F daily (β=0.50, 95%
CI: 0, 1.00) compared with those living at the shortest distance.
Moreover, living in or attending a school in a neighbourhood with
3 km RFE indices ≥2.5 was associated with up to a half serving less
of V&F (β=-0.40, 95% CI: -0.81, 0.005 for residential neighbourhood and β=-0.50, 95% CI: -0.91, -0.09 for school neighbourhood).
Table 3.
Covariate-adjusted Associations (OR and 95% CI) Between Measures of the School Neighbourhood Food Environment
and Dietary Outcomes in the QUALITY Study*
≥5 Servings of
V&F/Day
Beverages/Day
(n=489)
Eating/Snacking
Out
≥Once/Week
(n=502)
≥Once/Week
(n=502)
1.03 (0.63-1.68)
1.26 (0.77-2.06)
1
0.93 (0.56-1.55)
1.00 (0.62-1.62)
1
1.05 (0.67-1.65)
1.20 (0.79-1.81)
1
1.14 (0.70-1.86)
1.14 (0.73-1.78)
1
1.18 (0.66-2.10)
1.59 (0.95-2.64)‡
1
0.87 (0.51-1.48)
0.77 (0.48-1.23)
1
1.23 (0.79-1.94)
1.39 (0.89-2.17)
1
1.34 (0.84-2.14)
1.22 (0.77-1.93)
1
1.13 (0.66-1.91)
1.10 (0.68-1.81)
1
0.99 (0.58-1.68)
1.48 (0.91-2.39)
1
1.10 (0.69-1.77)
0.94 (0.61-1.47)
1
1.08 (0.68-1.71)
0.69 (0.43-1.10)
1
0.99 (0.55-1.78)
0.82 (0.49-1.35)
1
1.37 (0.74-2.51)
1.64 (0.96-2.79)‡
1
0.63 (0.37-1.06)‡
0.78 (0.48-1.28)
1
0.97 (0.56-1.67)
1.55 (0.91-2.64)
1
1.59 (0.85-2.94)
1.25 (0.69-2.25)
1
0.97 (0.54-1.75)
1.06 (0.64-1.76)
1
0.85 (0.50-1.47)
0.96 (0.58-1.57)
1
1.25 (0.71-2.20)
1.53 (0.93-2.50)‡
1
1.34 (0.69-2.60)
1.39 (0.80-2.41)
1
1.04 (0.56-1.93)
0.98 (0.59-1.61)
1
0.71 (0.38-1.35)
0.81 (0.47-1.41)
1
0.75 (0.41-1.35)
1.03 (0.61-1.73)
1
0.63 (0.39-1.04)‡
1
0.96 (0.60-1.51)
1
0.74 (0.47-1.15)
1
0.93 (0.61-1.41)
1
0.67 (0.41-1.08)‡
1
1.61 (1.01-2.56)†
1
0.83 (0.53-1.30)
1
1.25 (0.81-1.91)
1
(n=489)§
Proximity measures
Model 1 – Distance to nearest supermarket
Farthest (>1565 m)
Shortest (<670 m)
Model 2 – Distance to nearest fast-food restaurant
Farthest (>1460 m)
Shortest (<680 m)
Model 3 – Distance to nearest convenience store
Farthest (>834 m)
Shortest (<370 m)
Density measures
Model 4 – Density of supermarkets
Lowest
Intermediate
Highest
Model 5 – Density of fast-food restaurants
Lowest
Intermediate
Highest
Model 6 – Density of convenience stores
Lowest
Intermediate
Highest
Retail food environment measures
≥2 (21.9%)
<2 (78.1%)
≥2.5 (22.9%)
<2.5 (77.1%)
* Separate GEE (generalized estimating equations) model with logit link function for each main exposure and each outcome, adjusted for child’s age, sex,
parental education, household income, school neighbourhood material deprivation and school neighbourhood population density.
†p≤0.05; ‡p≤0.10.
§ When treated as a continuous outcome, 3 km RFE Index ≥2.5 (vs. less) was associated with V&F intake (Beta=-0.50, 95% CI: -0.91, -0.09).
OR=odds ratio; CI=confidence interval; V&F=vegetables and fruit; RFE=retail food environment.
DISCUSSION
We examined associations between indicators of neighbourhood
food environments and dietary outcomes among children with a
family history of obesity. The findings suggest that the availability
of fast-food restaurants and convenience stores in children’s neighbourhood environments may be associated with their intake of
V&F, and the likelihood of eating/snacking out and consuming
delivered/take-out foods. This extends recent research on built environments and children’s diets. Although associations tended to be
weak in magnitude, observed associations are overall consistent
with current research on obesogenic environments and health.
As previously reported,5-7 we found no consistent associations
between a greater availability of supermarkets and more favourable
dietary outcomes. Supermarkets typically offer a large variety of
healthful foods, including vegetables and fruits, at lower costs.25
However, there appear to be very few “food deserts” in Montreal, i.e.,
neighbourhoods where residents are considered to have poor access
to supermarkets.26 Associations between the availability of supermarkets and diets may be more likely to emerge in areas with less
equitable distributions of supermarkets and may be more relevant to
adult populations. In contrast to supermarket availability, we found
more evidence that the availability of fast-food restaurants and convenience stores was associated with children’s diets, particularly with
the likelihood of eating or having a snack in a food establishment.
These findings suggest that easy access to unhealthful foods may be
more of a concern than poor access to more healthful foods.27
Geographic clustering of fast-food restaurants and convenience
stores around schools has been described previously,14 although our
findings suggest that associations between access to these food
establishments and children’s diets were more consistent for residential than for school neighbourhood exposures. This may be
related to the relatively young age of participants; school neighbourhoods may become more important during adolescence, when
students attending secondary school are typically authorized to
leave school grounds.28
Use of the RFE indices revealed that children residing in or
attending a school in neighbourhoods with a preponderance of
unhealthful food establishments scored most poorly on dietary outcomes.29 An indicator of relative access to types of food establishments is a useful complement to proximity- and density-based
indicators, as commercial destinations tend to be geographically
clustered such that higher numbers of fast-food restaurants are
often associated with more supermarkets and fruit and vegetable
stores as well.
Restricting analyses to the subgroup of children living >1.5 km
from their school allowed us to partially distinguish associations
with residential neighbourhood environments from associations
with school neighbourhood environments. However, the results of
these subanalyses are likely not generalizable to the entire sample.
In this subgroup, children who lived farther away from their school
were more likely to be driven to or from their school than to travel by bus. A higher likelihood of car travel may lead to more oppor-
Table 4.
Covariate-adjusted Associations (OR and 95% CI) of Residential and School Neighbourhood Densities of Food
Establishments with Dietary Outcomes in Children Living More Than 1.5 km From Their School, QUALITY Study
≥5 Servings of
V&F/Day
Beverages/Day
(n=189)
Eating/Snacking
Out
≥Once/Week
(n=191)
1.01 (0.38-2.69)
0.79 (0.33-1.88)
1
1.00 (0.38-2.63)
1.32 (0.56-3.12)
1
0.66 (0.25-1.72)
0.70 (0.30-1.63)
1
1.27 (0.44-3.66)
2.13 (0.86-5.31)‡
1
0.91 (0.32-2.53)
0.67 (0.27-1.65)
1
1.69 (0.62-4.64)
1.22 (0.51-2.95)
1
0.67 (0.25-1.79)
0.79 (0.34-1.87)
1
3.45 (1.10-10.84)†
1.81 (0.70-4.73)
1
0.71 (0.24-2.08)
0.56 (0.21-1.45)
1
1.53 (0.53-4.40)
1.45 (0.57-3.69)
1
0.32 (0.11-0.93)†
0.41 (0.16-1.05)‡
1
1.78 (0.57-5.60)
1.67 (0.62-4.51)
1
1.70 (0.63-4.60)
1.06 (0.49-2.30
1
0.72 (0.28-1.88)
1.01 (0.45-2.26)
1
0.76 (0.31-1.90)
1.17 (0.54-2.53)
1
0.80 (0.26-2.48)
1.49 (0.60-3.71)
1
2.87 (1.16-7.10)†
0.87 (0.34-2.21)
1
1.04 (0.46-2.37)
1.31 (0.54-3.19)
1
1.63 (0.70-3.82)
1.76 (0.83-3.72)
1
1.51 (0.58-3.93)
2.84 (1.16-6.97)†
1
2.21 (0.76-6.47)
0.93 (0.37-2.37)
1
0.58 (0.20-1.68)
0.70 (0.25-1.92)
1
1.26 (0.47-3.35)
1.19 (0.53-2.69)
1
0.53 (0.16-1.81)
0.98 (0.32-3.04)
1
(n=189)
Residential environment*
Lowest
Intermediate
Highest
Lowest
Intermediate
Highest
Lowest
Intermediate
Highest
School environment**
Lowest
Intermediate
Highest
Lowest
Intermediate
Highest
Lowest
Intermediate
Highest
≥Once/Week
(n=191)
* Separate logistic regression models for each exposure and each outcome, adjusted for child’s age, sex, parental education, household income, residential
neighbourhood material deprivation and residential population density.
** Separate GEE (generalized estimating equations) models with logit link function for each exposure and each outcome, adjusted for child’s age, sex, parental
education, household income, school neighbourhood material deprivation and school neighbourhood population density.
†p≤0.05, ‡p≤0.10.
OR=odds ratio; CI=confidence interval; V&F=vegetables and fruit.
tunistic purchases by parents, including those at drive-through
restaurants, given the extended potential path area.9,30 This may,
in part, explain the higher fast-food intake among children living
farther away from fast-food restaurants.
Initiatives to create zones around schools with limited access to
fast-food restaurants and convenience stores have been proposed.31
Such initiatives may have a positive impact on children’s diet, particularly in the context of ecological interventions in which multiple levels of obesogenic environments are targeted. Although
school neighbourhoods might be more compelling targets, policies
to limit access to unhealthful food establishments in residential
neighbourhoods should be further investigated.
The strengths of this study include the use of a valid and reliable
method to measure children’s diet and the use of objective measures to characterize neighbourhood food environments. Overall,
the findings should be interpreted with caution, given the number
of associations tested and the increased risk of type-1 error. The
results should thus be seen as exploratory and in need of confirmation in future studies. Other limitations include the possibility
that children with certain dietary patterns were self-selected
through their parents to reside in neighbourhoods with particular
food establishment profiles. Moreover, because the majority of children lived within a short walking distance of their school, it was
not possible to distinguish entirely between the associations of residential vs. school neighbourhood environments with children’s
diets. While we used a GIS to quantify the availability of various
types of food establishment,32 others have used measures of perceived access.33 Parents and children may incorporate aspects other
than local availability to formulate perceptions of access, such as car
ownership, parental permissiveness and available pocket money;
this should be examined in future research that includes both GIS
and perceived measures. Last, since the children in this study were
relatively young (8-10 years), associations of interest may be mediated and/or confounded by parental diet; however, there were no
measures of parent diet in the QUALITY study. Maternal BMI was
considered as a proxy for mother’s diet, but was not retained
because its inclusion in the models did not change main exposure
coefficients substantively and because the study design required at
least one parent to be obese.
In conclusion, our findings suggest that among children aged
8-10 years, residential neighbourhood food environments are more
strongly associated with dietary outcomes than are school neighbourhood food environments. Although the magnitude of associations is relatively small, the potential to affect population dietary
behaviours and related health outcomes may be substantial. Frequent and widespread food purchasing opportunities within children’s environments may be one factor amenable to interventions
to improve diets.
REFERENCES
1.
2.
3.
4.
5.
Galvez MP, Pearl M, Yen IH. Childhood obesity and the built environment:
A review of the literature from 2008-2009. Curr Opin Pediatr 2010;22(2):2027.
Rahman T, Cushing RA, Jackson RJ. Contributions of built environment to
childhood obesity. Mt Sinai J Med 2011;78(1):49-57.
Dunton GF, Kaplan J, Wolch J, Jerrett M, Reynolds KD. Physical environmental correlates of childhood obesity: A systematic review. Obes Rev
2009;10(4):393-402.
Larson NI, Story MT, Nelson MC. Neighborhood environments. Disparities in
access to healthy foods in the U.S. Am J Prev Med 2009;36(1):74-81.
Pearce J, Hiscock R, Blakely T, Witten K. The contextual effects of neighbourhood access to supermarkets and convenience stores on individual fruit and
vegetable consumption. J Epidemiol Community Health 2008;62(3):198-201.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
Skidmore P, Welch A, van Sluijs E, Jones A, Harvey I, Harrison F, et al. Impact
of neighbourhood food environment on food consumption in children aged
9-10 years in the UK SPEEDY (Sport, Physical Activity and Eating behaviour:
Environmental Determinants in Young people) study. Public Health Nutr
2010;13(7):1022-30.
Timperio A, Ball K, Roberts R, Campbell K, Andrianopoulos N, Crawford D.
Children’s fruit and vegetable intake: Associations with the neighbourhood
food environment. Prev Med 2008;46(4):331-35.
Jennings A, Welch A, Jones AP, Harrison F, Bentham G, van Sluijs EM, et al.
Local food outlets, weight status, and dietary intake: Associations in children
aged 9-10 years. Am J Prev Med 2011;40(4):405-10.
Timperio AF, Ball K, Roberts R, Andrianopoulos N, Crawford D. Children’s
takeaway and fast-food intakes: Associations with the neighbourhood food
environment. Public Health Nutr 2009;12(10):1960-64.
An R, Sturm R. School and residential neighborhood food environment and
diet among California youth. Am J Prev Med 2012;42(2):129-35.
Casey R, Oppert J-M, Weber C, Charreire H, Salze P, Badariotti D, et al. Determinants of childhood obesity: What can we learn from built environment
studies? Food Quality and Preference 2011: doi:10.1016/j.foodqual.2011.06.003.
Papas MA, Alberg AJ, Ewing R, Helzlsouer KJ, Gary TL, Klassen AC. The built
environment and obesity. Epidemiol Rev 2007;29:129-43.
Gebauer H, Laska MN. Convenience stores surrounding urban schools: An
assessment of healthy food availability, advertising, and product placement.
J Urban Health 2011;88(4):616-22.
Kestens Y, Daniel M. Social inequalities in food exposure around schools in
an urban area. Am J Prev Med 2010;39(1):33-40.
Lambert M, van Hulst A, O’Loughlin J, Tremblay A, Barnett TA, Charron H,
et al. The Quebec Adipose and Lifestyle Investigation in Youth (QUALITY)
cohort. Int J Epidemiol 2011;Epub July 23.
Johnson RK, Driscoll P, Goran MI. Comparison of multiple-pass 24-hour recall
estimates of energy intake with total energy expenditure determined by doubly labeled water method in young children. J Am Diet Assoc
1996;96(11):1140-44.
Health Canada. Canadian Nutrient File. Available at: http://www.hc-sc.gc.ca/fnan/nutrition/fiche-nutri-data/cnf_downloads-telechargement_fcen-eng.php
(Accessed November 15, 2011).
Health Canada. Eating Well with Canada’s Food Guide. A Resource for Educators and Communicators. Available at: http://www.hc-sc.gc.ca/fn-an/foodguide-aliment/index-eng.php (Accessed November 15, 2011).
Paquet C, Daniel M, Kestens Y, Leger K, Gauvin L. Field validation of listings
of food stores and commercial physical activity establishments from secondary data. Int J Behav Nutr Phys Act 2008;5:58.
Carlos HA, Shi X, Sargent J, Tanski S, Berke EM. Density estimation and adaptive bandwidths: A primer for public health practitioners. Int J Health Geogr
2010;9:39.
Lebel A, Kestens Y, Pampalon R, Thériault M, Daniel M, Subramanian SV.
Local context influence, activity space, and foodscape exposure in two Canadian metropolitan settings: Is daily mobility exposure associated with overweight? J Obes 2012;2012:ID 912645. Epub 2011 Dec 28.
Spence JC, Cutumisu N, Edwards J, Raine KD, Smoyer-Tomic K. Relation
between local food environments and obesity among adults. BMC Public
Health 2009;9:192.
Pampalon R, Hamel D, Gamache P, Raymond G. A deprivation index for
health planning in Canada. Chronic Dis Can 2009;29(4):178-91.
Chaix B, Merlo J, Evans D, Leal C, Havard S. Neighbourhoods in ecoepidemiologic research: Delimiting personal exposure areas. A response to
Riva, Gauvin, Apparicio and Brodeur. Soc Sci Med 2009;69(9):1306-10.
Diez Roux AV, Mair C. Neighborhoods and health. Ann N Y Acad Sci
2010;1186:125-45.
Apparicio P, Cloutier M-S, Shearmur R. The case of Montreal’s missing food
deserts: Evaluation of accessibility to food supermarkets. Int J Health Geogr
2007;6(4).
Hackett A, Boddy L, Boothby J, Dummer TJB, Johnson B, Stratton G. Mapping
dietary habits may provide clues about the factors that determine food choice.
J Hum Nutr Diet 2008;21(5):428-37.
Laska MN, Hearst MO, Forsyth A, Pasch KE, Lytle L. Neighbourhood food
environments: Are they associated with adolescent dietary intake, food purchases and weight status? Public Health Nutr 2010;13(11):1757-63.
29. Babey SH, Wolstein J, Diamant A. Food environments near home and school
related to consumption of soda and fast food. Los Angeles, CA: UCLA Center
for Health Policy Research, 2011.
30. Páez A, Gertes Mercado R, Farber S, Morency C, Roorda M. Relative accessibility deprivation indicators for urban settings: Definitions and application to
food deserts in Montreal. Urban Studies 2012;47(7):1415-38.
31. Association pour la santé publique du Québec. The school zone and nutrition: Courses of action for the municipal sector. Available at:
http://www.aspq.org/documents/file/guide-zonage-version-finale-anglaise.pdf
(Accessed November 15, 2011).
32. Charreire H, Casey R, Salze P, Simon C, Chaix B, Banos A, et al. Measuring the
food environment using geographical information systems: A methodological review. Public Health Nutr 2010;13(11):1773-85.
33. Veugelers P, Sithole F, Zhang S, Muhajarine N. Neighborhood characteristics
in relation to diet, physical activity and overweight of Canadian children.
Int J Pediatr Obes 2008;3(3):152-59.
RÉSUMÉ
Objectifs : Rares sont les études du milieu bâti qui s’intéressent aux
environnements alimentaires des quartiers par rapport aux régimes
alimentaires des enfants. Nous avons examiné les associations entre
l’accès des quartiers résidentiels et scolaires à différents types
d’établissements alimentaires et les régimes des enfants.
Méthode : Nous avons analysé les données de l’étude QUALITY
(QUebec Adipose and Lifestyle InvesTigation in Youth), une étude en
cours sur l’histoire naturelle de l’obésité chez 630 jeunes Québécois de
8 à 10 ans ayant une histoire parentale d’obésité (n=512). Trois rappels
alimentaires de 24 heures ont servi à évaluer l’apport en fruits et légumes
et en boissons édulcorées au sucre. À l’aide de questionnaires, nous avons
déterminé la fréquence des repas et des collations pris à l’extérieur et la
consommation d’aliments livrés à domicile ou à emporter. Nous avons
caractérisé l’environnement alimentaire des quartiers résidentiels et
scolaires au moyen d’un système d’information géographique. Les
variables étaient la distance jusqu’au supermarché, au restaurant rapide
et au dépanneur le plus proche, et les densités de chacun de ces types
d’établissements, calculées sur un réseau tampon d’1 km autour du
domicile et de l’école de chaque enfant. Des indices d’environnement
alimentaire de détail ont aussi été calculés. La régression logistique
multivariée (accès à partir du domicile) et des équations d’estimation
généralisées (accès à partir de l’école) ont servi à l’analyse.
Résultats : L’accès des quartiers résidentiels et scolaires aux
supermarchés n’était pas associé aux régimes des enfants. Le fait
d’habiter un quartier où les restaurants rapides et les dépanneurs sont
moins accessibles était associé à une plus faible probabilité de prendre
des repas et des collations à l’extérieur. Les enfants qui fréquentaient des
écoles de quartiers comptant davantage d’établissements alimentaires
malsains que d’établissements sains ont obtenu les pires scores pour ce
qui est de leur régime.
Conclusions : Des enquêtes plus poussées sont nécessaires pour
formuler des politiques qui influencent les occasions d’achat d’aliments à
l’échelle des quartiers, particulièrement l’accès aux restaurants rapides et
aux dépanneurs.
Mots clés : milieu bâti; enfant; régime alimentaire; environnement
alimentaire; quartier résidentiel; quartier scolaire; cohorte QUALITY
Physical Activity and Nutrition Among Youth in Rural, Suburban
and Urban Neighbourhood Types
Cindy Shearer, PhD,1 Chris Blanchard, PhD,2 Sara Kirk, PhD,3 Renee Lyons, PhD,4 Trevor Dummer, PhD,5
Robert Pitter, PhD,6 Daniel Rainham, PhD,7 Laurene Rehman, PhD,3 Chris Shields, PhD,6 Meaghan Sim, MSc3
ABSTRACT
Objectives: Physical activity and nutrition are essential to healthy living and particularly important during youth, when growth and development are
key. This study examined rates of physical activity (PA) and diet quality (DQ) among youth in grades 7 to 9 in Halifax, Nova Scotia, during the 2008/09
school year and tested differences among students in rural, urban and suburban neighbourhood types of high and low socio-economic status (SES).
Methods: Youth in grades 7 through 9 (aged 12-16; 53% male) from six schools (N=380), stratified by neighbourhood type (urban, suburban, rural)
and SES, wore accelerometers for up to 7 days (mean=4.14, standard deviation=1.49) and completed a nutritional survey.
Results: The findings suggest important differences in PA and DQ across SES and neighbourhood type. Specifically, rates of moderate to vigorous
physical activity among youth from schools in lower socio-economic areas were higher in urban than in suburban or rural settings. Furthermore, DQ was
better among youth in higher than in lower socio-economic urban settings.
Conclusions: Understanding these differences in PA and DQ across rural, urban and suburban environments of high and low SES may highlight
subgroups and targeted geographic areas for the design of interventions to improve rates of PA and health nutrition.
Key words: Physical activity; nutrition; youth; built environment; socioeconomic status
P
romoting physical activity (PA) and diet quality (DQ) during
the adolescent years is particularly important, as research suggests that behaviours formed in adolescence extend into adulthood1,2 and carry consequences for long-term health. Past research
has revealed that less than 30% of 7th grade students in Nova Scotia
met the recommended level of PA (60 minutes per day, 5 days per
week) to achieve health benefits,3 and their diets did not meet
Canada’s Food Guide (e.g., more than 70% did not fulfill the daily
requirement of 6-8 servings [depending on age and sex] of fruits
and vegetables),4 making these two behaviours an important focus
for health promotion and intervention efforts.
Extant research has highlighted important socio-economic and
neighbourhood differences in PA and DQ. For instance, youth from
higher socio-economic backgrounds have been found to engage in
more PA5 and to have better DQ (diets that have variety, adequacy,
moderation and balance)6 than youth from lower socio-economic
backgrounds. However, the findings with regard to differences in
neighbourhood type (urban/suburban/rural) in PA and DQ are
somewhat mixed. Some studies of youth obesity have demonstrated higher rates of overweight and obesity in rural than urban and
suburban areas.7 Other studies have revealed lower rates of PA
among youth from urban neighbourhoods than those from rural
and/or suburban environments,8 whereas others have reported no
differences across neighbourhood types,9,10 and some have demonstrated higher rates of PA among urban than rural youth.11 Suburban contexts are often neglected in extant research. Suburban
neighbourhoods have characteristics of both rural and urban environments in that they are better connected to urban centres than
rural environments, yet they may sit at a distance from points of
interest that are not walkable.
With regard to DQ, Veugelers et al.12 reported a linkage between
rural environments and higher dietary fat and calorie consumption among Canadian children. Yet, convenience and “fast-food”
Author Affiliations
1. Atlantic Health Promotion Research Centre, Dalhousie University, Halifax, NS
2. Department of Medicine, Dalhousie University, Halifax, NS
3. School of Health and Human Performance, Faculty of Health Professions, Dalhousie
University, Halifax, NS
4. Bridgepoint Collaboratory for Research and Innovation, Bridgepoint Health,
University of Toronto, Toronto, ON
5. Population Cancer Research Program, Department of Pediatrics, Dalhousie
University, Halifax, NS
6. School of Recreation Management & Kinesiology, Acadia University, Wolfville, NS
7. Environmental Science Program, Faculty of Science, Dalhousie University, Halifax,
NS
Correspondence: Dr. Cindy Shearer, Atlantic Health Promotion Research Centre,
Dalhousie University, 1535 Dresden Row, Halifax, NS B3J 3T1, Tel: 902-494-2604,
Fax: 902-494-3594, E-mail: [email protected]
Acknowledgements: This research was supported by the Canadian Institutes of
Health Research’s Institute of Human Development, Child and Youth Health and
Institute of Nutrition, Metabolism and Diabetes; and the Heart and Stroke Foundation
of Canada, through the Built Environment, Obesity and Health Initiative. The authors
thank the ENACT team, including its principal investigators Renee Lyons and Jill Grant;
co-investigators (not listed as co-authors) Michael Arthur, Andrea Chircop,
Patricia Manuel and Louise Parker; community and policy partners Janet Barlow,
Diana Dibblee, Amy MacDonald, Roxane MacInnis, Michelle Murton, Clare O’Connor,
Paul Shakotko and Jacqueline Spiers; and staff and students Meredith Flannery,
Andrew Harding, Nicole Landry, Kathryn MacKay, Gillian McGinnis, Julie Rouette and
Stephanie Wood.
PHYSICAL ACTIVITY AND NUTRITION AMONG YOUTH
outlets – key features of obesogenic environments13,14 – are more
prevalent in urban areas. Thus, the urban environment may expose
youth to more unhealthy options and contribute to poor diet quality.15 The nutrition environment in the suburban context is also
unique in that the density of fast-food outlets may be much less
than in urban environments, but because of location outside the
urban core more frequent commuting may be required through
areas where unhealthy foods are readily available.
Better DQ has been shown in previous research to be associated
with improved health outcomes in adults,16 but there is a paucity of
data on DQ and health in children.17 A recent publication has
demonstrated an independent association between overall DQ and
academic performance in children,18 suggesting that improving DQ
may have impacts beyond health outcomes alone. Given that few
studies have considered the potential influence of socio-economic
status (SES) and neighbourhood type on PA and diet in youth,19 further research is needed to clarify their potential importance for this
population. Understanding differences in DQ and PA for subgroups
of this population is important for the development of interventions aimed at improving the health behaviours of youth.12,20
The purpose of the present study was to compare PA and DQ
among youth from schools of higher and lower SES in rural, suburban and urban neighbourhood types. It was hypothesized that
PA and DQ would be more favourable in higher versus lower SES
environments. It was also hypothesized that SES and neighbourhood type would interact when influencing PA.21
METHODS
institutional and recreational uses and a pattern of established neighbourhoods with low- to medium-density residential uses. Eligible
schools in each neighbourhood type were organized by SES and
divided into tertiles. One high and one low SES school was randomly selected from the higher and lower tertiles. Only one school that
was approached declined to participate.
Participating urban schools were located in areas that had high
residential density and street connectivity, high sidewalk availability, more mixed land uses and greater population density. Suburban
schools were located in areas with lower residential density and
street connectivity, and land uses that were spatially segregated.
Finally, rural schools were in areas that were automobile reliant,
with low residential density and street connectivity, no sidewalks,
and schools placed far from residential land uses.
Recruitment took place in one school at a time during the 2008
and 2009 school years. Students were recruited through presentations in each 7th to 9th grade classroom. Information packages,
including consent forms, were distributed to obtain parental consent; 27% of these forms were returned for participation in the
study. In addition to completing surveys of dietary intake and
health behaviours, students were asked to wear an accelerometer
and GPS (Global Positioning System) device (to measure their
geospatial footprint, which is not a focus of the current report) for
a period of one week. All participants were entered in a prize draw
for a gift card for participating. Furthermore, cash incentives were
provided to encourage participants to wear the equipment ($20 for
6 or fewer days of wear, $30 for 7 days). All surveys (diet and health
behaviours) and measurements (height/weight) were collected prior
to distributing the accelerometers to students.
Recruitment
This research protocol was approved by the principal investigators’
institutional review board as well as the review panel in place at
the school board from which schools were recruited. Schools in the
Halifax Regional Municipality, NS, were eligible for inclusion if they
1) enrolled students in Grades 7 through 9, and 2) did not offer a
French immersion program (as these schools draw a greater proportion of students from areas outside the school’s eligible neighbourhood). Within the Halifax Regional School Board, 38 schools
fit these criteria; 5 were located in rural areas, 24 in suburban areas
and 9 in urban areas. Six public schools were stratified by schoollevel SES and neighbourhood type.
School-level SES was determined by the median household income
of the school’s census dissemination area (a term used by Statistics
Canada to refer to a small area composed of one or more neighbouring blocks, with a population of 400 to 700 persons), based on
2006 census data. Urban, suburban and rural categories were designated through a two-step process. The first step distinguished rural
from urban using the Statistics Canada population-based definition
of an urban area.* The Halifax Regional Municipality municipal planning guidelines† were then used to subdivide urban areas into urban
and suburban categories according to urban development patterns,
including housing density and a mix of commercial, institutional
and recreational uses. Urban areas had a mix of high-density residential, commercial, institutional and recreational uses, whereas suburban areas had a mix of low- and medium-density commercial,
* >1,000 persons per km2; see http://www.statcan.gc.ca/pub/92f0138m/
92f0138m2008001-eng.pdf
† See http://www.halifax.ca/districts/dist17/documents/RegionalPlan.pdf
Measures
Demographic Features
Students reported their ethnicity (identified from a list of options,
including “other” and space to describe) and sex within the survey
of health behaviours and dietary intake.
Body Mass Index (BMI)
Students had their weight and height measured by trained research
assistants in a private area of the school. These measurements were
used to calculate BMI.
Physical Activity
PA was assessed objectively using the Actigraph GT1M (Actigraph:
Pensacola, FL) accelerometer (placed on the right hip) for seven
consecutive days. This accelerometer has documented evidence of
concurrent validity and inter-instrument reliability in several studies of children and adolescents.22 Students were asked to wear the
device for all waking hours of the day and to remove it for waterbased activities and contact sports. At least one valid day (i.e., ≥8
hours of valid data) was required to be included in analyses; 92.7%
of those who wore an accelerometer met this requirement. Students
averaged 4.14 (SD=1.49) days of valid data. Raw accelerometer
counts were converted to minutes per day of moderate (i.e., activities that cause youth to sweat a bit and breathe harder) to vigorous
(i.e., activities that cause youth to sweat and be out of breath) physical activity (MVPA) using age-specific count thresholds developed
by Freedson and colleagues.23
Table 1.
Sample Characteristics by School-level Socio-economic Status and Urban, Suburban and Rural Built Environments
(N=380)
Demographic Characteristics
Median income of census dissemination area, $
Season of data collection
Age, %
<12
13
14
15
16
Male, %
Grade, %
7
8
9
White, %
Mean (SD) BMI
High SES (School-level)
Urban
Suburban
Rural
n=45
n=25
n=4
54,827
62,834
50,325
Oct/Nov
May/Jun
May/Jun
Low SES (School-level)
Urban
Suburban
Rural
n=63
n=97
n=86
26,641
47,821
30,527
Nov/Dec
Mar/Apr
May/Jun
53.3
40.0
6.7
0.0
0.0
60.0
12.0
56.0
16.0
16.0
0.0
52.0
20.3
35.9
32.8
9.4
1.6
45.3
31.7
36.5
25.4
6.3
0.0
52.4
14.4
36.1
37.1
11.3
1.0
55.7
11.6
33.7
38.4
15.1
1.2
52.3
55.6
40.0
4.4
88.9
19.0 (2.9)
64.0
24.0
12.0
80.0
20.8 (3.6)
34.4
34.4
31.3
95.3
22.5 (4.5)
39.7
39.7
20.6
60.3
21.9 (4.5)
33.0
35.1
32.0
85.6
22.8 (5.4)
33.7
39.5
26.7
93.0
22.5 (4.8)
SES=socio-economic status (school-level).
Table 2.
Moderate-to-Vigorous Physical Activity (MVPA) Per
Day and Diet Quality by School-level Socioeconomic Status and Urban, Suburban and Rural
Built Environments
Minutes of MVPA/Day
Mean (SD)*
Median
High SES
Urban
Suburban
Rural
Low SES
Urban
Suburban
Rural
Diet Quality
Mean (SD)*
Median
71.52 (23.00)
69.62 (41.90)
48.39 (32.41)
69.83
60.08
41.23
67.29 (8.12)†
65.12 (8.27)
63.98 (8.02)
69.76
63.96
64.84
82.50 (39.13)†
55.00 (31.03)‡
58.81 (30.63)‡
73.13
48.00
49.83
59.75 (7.06)‡
63.02 (7.60)
64.20 (7.67)
59.28
63.55
64.34
* Means with different superscripts were significantly different, p<0.05.
SES=socio-economic status (school-level).
RESULTS
Demographic data are presented in Table 1. Of the 380 students
recruited, 53% were male and 84% were white. Participant age
ranged from 12 or less to 16 years, with somewhat greater participation among younger students. This was reflected in the distribution by grade, which revealed lower participation by grade 9
students at most schools.
Preliminary analyses showed that MVPA per day was significantly
related to sex (F[1,344]=23.64, p=0.00), grade (F[1,344]=5.34,
p=0.01) and BMI (r=-0.16, p=0.00), whereas DQ differed by ethnicity (F[1,330]=6.02, p=0.02). Therefore, these were controlled for
in subsequent analyses.
Physical activity
Diet Quality
Nutritional intake was assessed by means of the Harvard
Youth/Adolescent Questionnaire (YAQ),24 a validated food frequency instrument suitable for adolescents in this age group. As
DQ is best represented in a composite measure, data from the YAQ
were used to calculate a Diet Quality Index (DQI)24 for each student. DQI values encompass dietary variety (i.e., overall variety and
variety within protein sources, to assess whether intake comes from
diverse sources both across and within food groups), adequacy
(i.e., the intake of dietary elements that must be supplied sufficiently
to guarantee a healthy diet), moderation (i.e., intake of food and
nutrients that are related to chronic diseases and that may need
restriction) and balance (i.e., the overall balance of diet in terms of
proportionality in energy sources and fatty acid composition).
Scores ranged from 0 to 100, the higher scores reflecting better DQ.
The DQI has been useful in cross-national comparisons of diet quality25 and has demonstrated important associations with other measures of healthy eating.18
Analytical plan
Demographic and descriptive statistics were generated, and a series
of zero-order correlations and between-subject ANOVAs were conducted to identify potential covariates (sex, grade, ethnicity and
BMI) for the main analyses. Once identified, a series of 2 (SES: high
vs. low) × 3 (neighbourhood type: urban, suburban, rural) analyses
of covariance (ANCOVA) were conducted on the MVPA per day and
DQ variables.
With regard to MVPA per day, the ANCOVA showed a significant
school-level SES × neighbourhood type interaction (F[2,343]=4.56,
p=0.01) (see Table 2). Follow-up between-subject ANCOVAs were
conducted for high and low SES groups separately and showed that
the effect of neighbourhood type was only present for low SES
schools (F[2,225]=14.49, p=0.00). Least significant difference posthoc analyses showed that low SES urban students engaged in significantly more MVPA per day than low SES suburban (p=0.00) and
rural (p=0.00) students; however, MVPA per day was similar for low
SES suburban and rural students (p=0.71).
Diet quality
With respect to DQ, the ANCOVA showed a significant school level
SES × neighbourhood type interaction (F[2,330]=4.21, p=0.02) (see
Table 2). Follow-up between-subject ANCOVAs were conducted for
urban, suburban and rural groups separately and showed that high
SES urban students had significantly better DQ scores than their
low SES counterparts (F[1,85]=14.41, p=0.00); however, there were
no school-level SES differences for suburban or rural students.
DISCUSSION
Physical activity
The health benefits of PA are well known, yet PA rates tend to
decline during the adolescent period.26 Understanding how PA rates
vary for subgroups during this phase of the lifespan can highlight
possible avenues for targeted interventions to improve youth
health. Consistent with our hypotheses, analyses revealed an interaction between SES and neighbourhood type. Higher rates of PA
were found among youth who attended school in a low SES urban
setting than among those who attended school in low SES rural or
suburban settings. The rural schools in our study had larger catchment areas and required bussing of most or all students, whereas
bussing was less common in urban schools. Therefore, active transportation to school may account for some of the difference in
MVPA in the urban and rural settings. Otherwise, there were few
differences noted among the low SES schools in their physical education programming or the availability of facilities for use outside
of school time (e.g., all low-SES schools had a gymnasium with
change rooms on site as well as outdoor paved areas for PA, fields
and running tracks available to them either on or off school
grounds; all reported that students would have the option to participate in intramural programs that involved physical activity five
days per week).
This finding is consistent with previous research of adult PA,
which has typically found higher rates of PA in urban than rural
areas.27 However, it is inconsistent with some research on children
and youth. In a review of studies examining the PA levels of children living in different built environments, Sandercock et al.8 found
that most studies either reported no difference among urban, rural
and (when examined) suburban environments or that children in
rural and/or suburban environments had higher PA levels than
those in urban environments. Most of the studies reviewed by
Sandercock et al.8 employed either self- or parent-reports of
child/youth PA, which are vulnerable to overreporting, social desirability influences and difficulty in recall.28 Our study improves on
the literature by using objectively measured PA across three distinct
neighbourhood types. Furthermore, greater PA among rural children is often attributed to outdoor play, which is more likely to
occur in younger age groups than the one studied here.29
Typically, explanations for rural-urban differences in adult PA
highlight the limited availability and accessibility of venues for
leisure-time PA and poor walkability of rural areas.30,31 This disparity is likely exacerbated when financial resources are lower. Indeed,
Parks et al.21 found that rural, lower-income adults were less than
half as likely as suburban, higher-income adults to meet PA recommendations. Canadian data indicate that parents from smaller
communities are less likely to report the availability of public and
private opportunities for PA and less likely to report that that those
opportunities meet their children’s needs;32 the data also indicate
that youth perceive a lack of opportunities close to home as a barrier to physical activity.33 Lower socio-economic regions are even
less likely to have venues for leisure-time PA than higher socioeconomic regions, and where they do exist the limited financial
resources of residents may preclude their use.34 In a qualitative
examination, low SES Canadian youth were more likely than their
high SES counterparts to report the proximity and cost of facilities
as factors that determined their participation in PA.35
The inclusion of a suburban comparison category represents a
further novel aspect of the current study, as few studies have gone
beyond the examination of simple urban/rural differences.8 A pattern of higher PA levels in children in suburban environments has
emerged when this category is considered.36 In our study, however, rates of youth PA were similar in suburban and rural settings
regardless of neighbourhood SES. This discrepancy may be rooted
in differences in measurement (our study measured PA objectively
by accelerometers, whereas Springer et al.36 employed self-report
measures). Developmental differences may also play a role, in that
Springer et al. studied high school students. These older adolescents
may have the autonomy to drive to nearby centres for PA, whereas younger adolescents may not. Suburban environments, by definition, are located outside of the urban core. Therefore,
opportunities for physical activities of interest to this age group (at
facilities such as rinks, skate parks and recreation centres) may not
be within walkable distance in either rural or suburban neighbourhoods.
Diet quality
Consistent with our hypothesis, DQ was found to be poorer among
youth who attended school in the low SES setting than among their
counterparts who attended school in the high SES setting. However, this pattern was found only for urban schools. Differences in
the nutrition environment and programming at these urban
schools were few: the high SES school had a vending machine for
drinks whereas the low SES school did not, and the low SES school
offered a breakfast program whereas the high SES school did not.
Yet, neither of these offers a clear explanation for the pattern of
findings that emerged. Socio-economic differences in DQ have
been highlighted quite consistently in the literature: 80% of articles
reviewed by Hanson and Chen5 revealed an association between
higher SES and greater DQ. Explanations for the association
between low SES and poor nutrition often highlight unsafe or
impoverished living environments with limited access to healthy
foods and/or limited knowledge of healthy eating practices.5 Further exploration of this phenomenon in this sample is ongoing,
with preliminary qualitative analysis suggesting that accessibility of
healthy foods plays an important role in food choice (data not
shown).
Much less research has focused on neighbourhood-type differences in DQ. There is some indication that dietary fat and calorie
consumption is higher among rural youth and families.12,37 Yet,
research examining the impact of the built environment features
on youth health has suggested that the accessibility of fast-food
restaurants, which is greater in urban than rural environments, is
an important predictor of obesity.15 Although the current study did
not find differences across neighbourhood type, the low SES urban
environment emerged as a setting with particularly poor DQ. Perhaps by combining the economic and geographic accessibility of
unhealthy foods (especially fast food), urban environments in
lower socio-economic areas may be particularly obesogenic.38
CONCLUSION
Several limitations need to be considered in interpreting the findings of the current study. As in all correlational research, it is important to acknowledge that the relations between neighbourhood
type, SES, PA and DQ presented here are not causal. Thus, we cannot conclude that living in a low SES urban environment causes
youth to engage in greater PA or to eat foods that result in a lower
DQ. Self-selection of individuals and families into particular neighbourhoods may play an important, albeit immeasurable, role in our
findings. If neighbourhood self-selection could be taken into
account, differences found between the six settings might be attenuated. Further self-selection into the study is another limitation.
Although the schools were randomly selected from identified strata, only one school was chosen to represent each stratum, limiting
the generalizability of our findings. Furthermore, the recruitment
rate was somewhat low, likely because of the greater extent of participant involvement (e.g., wearing and charging study equipment)
than in other studies, and grade 9 students may be somewhat
under-represented in comparison to students in grades 7 and 8.
Also, because schools were recruited one at a time, seasonal changes
may have affected different rates of PA across the six schools. With
regard to measurement, because accelerometers were required to
be removed during water and contact sports, the measure of MVPA
did not include these activities. Finally, because household income
was not self-reported, variations among students within the same
school area are not taken into account in these analyses.*
Despite these limitations, the current study contributes to an
improved understanding of the variation in PA and diet in adolescents from more or less urbanized neighbourhoods in several ways.
First, by employing accelerometers to measure PA, this paper
improves upon earlier descriptive work that has used primarily selfreport indices. Second, investigations rarely consider both PA and
dietary intake – i.e., both sides of the energy balance equation –
within the same study,19 which is critical to a greater understanding of the role of the built environment in obesity and other chronic diseases. This work provides detailed descriptive information on
youth PA levels and dietary intake for concurrent consideration and
reveals important differences in the patterning of these healthrelated behaviours across school-level income levels and neighbourhood types. Finally, studies that consider neighbourhood types
often include only rural and urban categories. By considering the
suburban environment, this study advances the current understanding of health behaviours among youth in these geographic
areas. It suggests that neighbourhood type and SES interact and
should both, therefore, be carefully considered in identifying both
areas of risk (e.g., rural vs. urban areas) and target behaviours
(e.g., diet quality vs. physical activity) in the development of initiatives aimed to promote PA and DQ among youth.
REFERENCES
1.
2.
3.
4.
5.
6.
7.
Kelder SH, Perry CL, Klepp KI, Lytle LL. Longitudinal tracking of adolescent
smoking, physical activity, and food choice behaviors. Am J Public Health
1994;84:1121-26.
Larson NI, Neumark-Sztainer D, Hannan PJ, Story M. Family meals during
adolescence are associated with higher diet quality and healthful meal patterns during young adulthood. J Am Diet Assoc 2007;107:1502-10.
Thompson AM, McHugh TL, Blanchard C, Campagna PD, Durant MA,
Rehman LA, et al. Physical activity of children and youth in Nova Scotia from
2001/02 and 2005/06. Prev Med 2009;49(5):407-9.
Campagna P, Amero M, Arthur M, Durant M, Murphy R, Porter J, et al. PACY
2005: Physical Activity Levels and Dietary Intake of Children and Youth in the
Province of Nova Scotia. Halifax, NS: Nova Scotia Department of Health Promotion and Protection and Department of Education, 2005.
Hanson MD, Chen E. Socioeconomic status and health behaviors in adolescence: A review of the literature. J Behav Med 2007;30:263-85.
Veugelers PJ, Fitzgerald AL, Johnston E. Dietary intake and risk factors for
poor diet quality among children in Nova Scotia. Can J Public Health
2005;96:212-16.
Ismailov RM, Leatherdale ST. Rural-urban differences in overweight and obesity among a large sample of adolescents in Ontario. Int J Pediatr Obes
2010;5:351-60.
* A separate analysis not included in the current article controlled for homelevel SES using a proxy measure that included measures of income, education and unemployment based on the census dissemination area of
participants’ home address. The inclusion of this control variable had no
impact on our findings but complicated the report and thus was omitted.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
Sandercock G, Angus C, Barton J. Physical activity levels of children living in
different built environments. Prev Med 2010;50:193-98.
Hodgkin E, Hamlin MJ, Ross JJ, Peters F. Obesity, energy intake and physical
activity in rural and urban New Zealand children. Rural Remote Health
2010;10:1336 (online).
Loucaides CA, Plotnikoff RC, Bercovitz K. Differences in the correlates of
physical activity between urban and rural Canadian youth. J Sch Health
2007;77:164-70.
Huang S, Hung W, Sharpe PA, Wai JP. Neighborhood environment and physical activity among urban and rural schoolchildren in Taiwan. Health Place
2010;16:470-76.
Veugelers P, Sithole F, Zhang S, Muhajarine N. Neighbourhood characteristics in relation to diet, physical activity and overweight of Canadian children. Int J Pediatr Obes 2008;3:153-59.
Battle EK, Brownell KD. Confronting the rising tide of eating disorders and
obesity: Treatment vs. prevention and policy. Addict Behav 1996;21:755-65.
Ebbeling CB, Pawlak DB, Ludwig DS. Childhood obesity: Public health crisis,
common sense cure. Lancet 2002;360:473-82.
Davis B, Carpenter C. Proximity of fast-food restaurants to schools and adolescent obesity. Am J Public Health 2009;99:505-10.
Kant AK. Dietary patterns and health outcomes. J Am Diet Assoc 2004;104:615-35.
Smithers LG, Golley RK, Brazionis L, Lynch JW. Characterizing whole diets of
young children from developed countries and the association between diet
and health: A systematic review. Nutr Rev 2011;69:449-67.
Florence MD, Asbridge M, Veugelers PJ. Diet quality and academic performance. J Sch Health 2008;78:209-15.
Kirk SFL, Penney TL, McHugh TLF. Characterizing the obesogenic environment: The state of the evidence with directions for future research. Obes Rev
2009;11:109-17.
Hoeslcher DN, Evans A, Parcel G, Kelder SH. Designing effective nutrition
interventions for adolescents. J Am Diet Assoc 2002;102:S52-S63.
Parks SE, Housemann RA, Brownson RC. Differential correlates of physical
activity in urban and rural adults of various socioeconomic backgrounds in
the United States. J Epidemiol Community Health 2003;57:29-35.
Trost SF. Measurement of physical activity in children and adolescents. Am J
Life Med 2007;1:299-314.
Freedson PS, Sirard J, Debold EP, Pate R, Dowda M. Calibration of the Computer Science and Applications, Inc. (CSA) accelerometer. Med Sci Sports Exerc
1997;29:S45.
Rockett RH, Breitenback M, Frazier L, Witschi J, Wolf AM, Field AE, et al.
Validation of a youth/adolescent food frequency questionnaire. Prev Med
1997;26:808-16.
Kim S, Haines PS, Siega-Riz AM, Popkin BM. The Diet Quality IndexInternational (DQI-I) provides an effective tool for cross-national comparison of diet quality as illustrated by China and the United States. J Nutr
2003;133:3476-84.
Irving HM, Adlaf EM, Allison KR, Paglia A, Dwyer JJ, Goodman J. Trends in
vigorous physical activity participation among Ontario adolescents, 19972001. Can J Public Health 2003;9:272-74.
King AC, Castro C, Wilcox S, Eyler AA, Sallis JF, Brownson RC. Personal and
environmental factors associated with physical inactivity among different
racial-ethnic groups of US middle-aged and older aged adults. Health Psychol
2000;19:354-64.
Sallis JF, Saelens BE. Assessment of physical activity by self-report: Status, limitations, and future directions. Res Q Exerc Sport 2000;71(2):1-14.
Joens-Matre RW, Welk GJ, Calabro MA, Russell DW, Nicklay E, Hensley LD.
Rural-urban differences in physical activity, physical fitness, and overweight
prevalence of children. J Rural Health 2008;24:49-54.
Yousefian A, Ziller E, Swarz J, Hartley D. Active living for rural youth: Addressing physical inactivity in rural communities. J Public Health Manag Pract
2009;15:223-31.
Frost SS, Goins RT, Hunter RH, Hooker SP, Bryant LL, Kruger J, Pluto D. Effects
of the built environment on activity of adults living in rural settings.
Am J Health Promot 2010;24:267-83.
Canadian Fitness and Lifestyle Research Institute. Local opportunities to be
active. Physical Activity Monitor 2005;Bulletin 5:64-78.
Walia S, Leipert B. Perceived facilitators and barriers to physical activity for
rural youth: An exploratory study using photovoice. Rural Remote Health
2012;12;1842. Epub.
Moore LV, DiezRoux AV, Evenson KR, McGinn AP, Brines SJ. Availability of
recreational resources in minority and low socioeconomic status areas.
Am J Prev Med 2008;34:16-22.
Humbert ML, Chad KE, Spink KS, Muhajarine N, Anderson KD, Bruner MW,
et al. Qual Health Res 2006;16:467-83.
Springer A, Hoelscher DM, Kelder SH. Prevalence of physical activity and
sedentary behaviors in US high school students by metropolitan status and
geographic region. J Phys Act Health 2006;3:365-80.
Lutz SM, Blaylock JR, Smallwood DM. Household characteristics affect food
choices. Food Rev 1993;16(2):12-17.
Burdette HL, Whitaker RC. Neighborhood playgrounds, fast food restaurants,
and crime: Relationships to overweight in low-income preschool children.
Prev Med 2004;38:57-63.
RÉSUMÉ
Objectifs : L’activité physique et la nutrition sont essentiels à un mode
de vie sain et particulièrement importantes durant la jeunesse, en pleine
période de croissance et de développement. Nous avons examiné les
taux d’activité physique (AP) et la qualité du régime (QR) d’élèves de la
7e à la 9e année à Halifax (Nouvelle-Écosse) durant l’année scolaire 20082009 et évalué les écarts entre les élèves des quartiers ruraux, urbains et
suburbains, de statut socioéconomique (SSE) faible et élevé.
Méthode : Des jeunes de la 7e à la 9e année (de 12 à 16 ans; 53 % de
garçons) fréquentant six écoles (N=380) stratifiées selon le type de
quartier (urbain, suburbain, rural) et le SSE ont porté des accéléromètres
pendant 7 jours ou moins (4,14 jours en moyenne, écart type de 1,49) et
rempli un questionnaire sur la nutrition.
Résultats : Les constatations de l’étude laissent entrevoir d’importants
écarts dans l’AP et la QR selon le SSE et le type de quartier. En particulier,
les taux d’activité physique modérée à vigoureuse chez les jeunes des
écoles de zones socioéconomiquement faibles étaient plus élevés en
milieu urbain qu’en milieu rural ou suburbain. De plus, la QR était
meilleure chez les jeunes des quartiers urbains de SSE élevé plutôt que
faible.
Conclusion : La connaissance de ces écarts dans l’AP et la QR entre les
environnements ruraux, urbains et suburbains, de SSE élevé ou faible,
peut faire ressortir des sous-groupes et des zones géographiques à cibler
pour concevoir des interventions qui améliorent les taux d’AP et de saine
alimentation.
Mots clés : activité physique; nutrition; jeunesse; milieu bâti; statut
socioéconomique
Creating Neighbourhood Groupings Based on Built Environment
Features to Facilitate Health Promotion Activities
Donald Schopflocher, PhD,1 Eric VanSpronsen, MPH,1 John C. Spence, PhD,2 Helen Vallianatos, PhD,3
Kim D. Raine, PhD,1 Ronald C. Plotnikoff, PhD,4 Candace I.J. Nykiforuk, PhD1
ABSTRACT
Objectives: Detailed assessments of the built environment often resist data reduction and summarization. This project sought to develop a method of
reducing built environment data to an extent that they can be effectively communicated to researchers and community stakeholders. We aim to help in
an understanding of how these data can be used to create neighbourhood groupings based on built environment characteristics and how the process of
discussing these neighbourhoods with community stakeholders can result in the development of community-informed health promotion interventions.
Methods: We used the Irvine Minnesota Inventory (IMI) to assess 296 segments of a semi-rural community in Alberta. Expert raters “created”
neighbourhoods by examining the data. Then, a consensus grouping was developed using cluster analysis, and the number of IMI variables to
characterize the neighbourhoods was reduced by multiple discriminant function analysis.
Results: The 296 segments were reduced to a consensus set of 10 neighbourhoods, which could be separated from each other by 9 functions
constructed from 24 IMI variables. Biplots of these functions were an effective means of summarizing and presenting the results of the community
assessment, and stimulated community action.
Conclusions: It is possible to use principled quantitative methods to reduce large amounts of information about the built environment into meaningful
summaries. These summaries, or built environment neighbourhoods, were useful in catalyzing action with community stakeholders and led to the
development of health-promoting built environment interventions.
Key words: Built environment; quantitative methods; health promotion; knowledge exchange; obesity reduction
I
n Canada, chronic diseases as leading causes of death are linked
by common preventable risk factors related to lifestyle: tobacco
use, unhealthy diet and physical inactivity.1 Recent research suggests that prevention efforts should target these risk factors and
environmental, economic, social and behavioural determinants of
health.1-3 Specifically, the built environment (BE) is implicated as an
important consideration for interventions focused on risk factors as
well as health determinants.
The relevance of the BE to prevention and reduction of adverse
health outcomes associated with chronic disease is clear in research
on obesogenic environments4-9 and on physical activity.10-13
Research on obesogenic environments is predicated on the idea
that obesity is a normal response to an abnormal environment and
that understanding, measuring and altering that environment is
central to intervention success.5 This is consistent with the settingsbased strategy of the Ottawa Charter, calling for environments that
make the healthy choice the easy choice.14 Within every community there is interaction between individuals, micro- and macroenvironments6 and types of environment (physical, economic,
social and political). These interactions shape what (healthy) choices
are available. It follows that examination of the environment
should facilitate identification of factors amenable to intervention.
Yet there is no consensus on how to measure environments or on
specific factors that might be changed to improve health. This
quandary is exacerbated in small or rural municipalities that do not
have administratively defined neighbourhoods, which typically
form the basis for measurement in larger urban areas.
Recent “state of the science” articles provide critical analyses of
existing measures for documenting the impact of BE on physical
activity and healthy eating.15,16 These reviews indicate that the field
has not progressed beyond simple description of variables and their
associations. Few analysis techniques have been identified to aggregate or to summarize these measures and thereby stimulate theory
construction. The Irvine-Minnesota Inventory (IMI) provides an
example; it is a comprehensive measure of macro- and micro-level
BE characteristics that may be linked to physical activity.17 The IMI
seeks ratings of 164 characteristics for each road segment (two facing sides of one street block) in a given setting. At the time of data
analysis for the current paper, there were no published methods
for summarizing the very large amount of data that result from use
of the IMI.
Author Affiliations
Edmonton, AB
2. Faculty of Physical Education and Recreation, University of Alberta, Edmonton, AB
3. Department of Anthropology, Faculty of Arts, University of Alberta, Edmonton, AB
4. School of Education, University of Newcastle, NSW, Australia
Correspondence: Dr. Candace Nykiforuk, 3-300 ECHA, 11405-87 Ave., Edmonton, AB
T6G 1C9, Tel: 780-492-4109, Fax: 780-492-0364, E-mail: [email protected]
Acknowledgements: Funding for this project was provided to C. Nykiforuk by
grants from the Heart and Stroke Foundation of Canada in partnership with the
Canadian Institutes of Health Research (CIHR). K. Raine and R. Plotnikoff are supported
by the CIHR Applied Research Public Health Chair Program. Raine’s Chair is funded
by the Heart and Stroke Foundation of Canada. We thank Laura Nieuwendyk for
conducting the community assessment and our community partners for their
participation and support.
NEIGHBOURHOOD GROUPINGS FOR HEALTH PROMOTION
This paper aims to 1) present a quantitative method for summarizing the information from comprehensive environmental inventories such as the IMI, and 2) demonstrate how these summaries
can contribute to dialogues with communities for the development
of health promotion interventions.
Figure 1.
Latitude and longitude plot of sidewalk presence on
the 296 segments observed in the town of
Bonnyville
METHODS
Background
This paper reports on a subproject of the Community Health and
the Built Environment (CHBE) project, which worked in partnership with communities to promote physical activity and healthy
eating by identifying and overcoming barriers in the built environment.18 The CHBE project included expert assessment of the BE,
followed participatory research principles,19 and employed active
and passive knowledge exchange20 to ensure that findings and
interventions were meaningful for public health practice and policy at community and regional levels. BE groupings were undertaken for this community because it did not have administratively
defined neighbourhoods to support characterization of the BE
analyses. The data presented here were collected in the town of
Bonnyville, a semi-rural community located in Alberta; the town’s
population was 5,832, within a municipal district population of
10,194.21
Ethical clearance for the project was received from the Health
Research Ethics Board (Panel B), University of Alberta.
Rating procedure
In the summer and fall of 2008, three community observers were
trained to use an adapted version of the IMI (which included an
additional 30 variables) to document 3,786 segments in four Alberta communities (a full description of tool adaptations is available
elsewhere18). In Bonnyville, all segments (n=296) were rated by one
observer. Ratings were registered on a Motorola MC35 handheld
computer running CyberTrack (v3.129) software. A Global Positioning System (GPS) reading was taken at the midpoint of each
segment.
different symbol, and a legend was provided for each plot. The set
of 197 plots (representing all original IMI and additional variables)
was presented to the six raters. Figure 1 shows an example plot.
Here, the four symbols differentiate the 296 segments on the basis
of having “No Sidewalks”, a sidewalk on one side of the street, or
having sidewalks on “Both Sides” of the street.
Adapting a sorting method used in multivariate research22 to the
current task, we asked each rater to use the information from the
complete set of plots to form a number of mutually exclusive and
jointly exhaustive groupings of segments in Bonnyville so that each
grouping represented a relatively homogenous region. No particular number of groupings was requested. Each rater was provided
with a number of plots marked only by the GPS coordinates on
which to draw their provisional and final groupings. The raters consisted of the observer of the Bonnyville segments and five other
members of the multidisciplinary research team.
Phase two: Consensus groupings
Analysis
A method of analysis was developed to reduce the information
from the IMI ratings to manageable proportions and allow it to be
effectively communicated. In the first phase, six expert raters were
presented with representations of IMI data and then requested to
use this information to create a set of groupings of geographically
contiguous segments by assigning each segment to a single group
(or “neighbourhood” based on BE characteristics). In the second
phase, these experts’ groupings were themselves analyzed by cluster analysis to form a single consensus grouping. In the third phase,
a discriminant function analysis was performed on the consensus
groupings using the items of the IMI to form meaningful scales of
items that separated the groupings (BE neighbourhoods).
Phase one: Rater groupings
Since a GPS reading had been taken at each segment, it was possible to plot data from all 296 segments for each IMI variable on a
separate two-dimensional longitude and latitude plot. Thus the spatial relations between segments were maintained. For variables with
multiple response categories, each category was represented by a
Each group for each rater was translated into a binary variable indicating which segments were present in the grouping. Each rater
provided either eight or nine groupings, and this resulted in a total
of 51 separate grouping vectors. These 51 vectors were the variables
used to form a consensus grouping. A hierarchical cluster analysis
using Ward’s method on squared Euclidean distances23 was conducted with SPSSv15. Several solutions were examined, and ultimately the solution with 10 groups was chosen for further analysis.
This procedure reduced the number of rated areas to be considered
from 296 to 10.
Phase three: Discriminant function analysis
The IMI variables were then examined to determine whether they
were suitable for inclusion in a discriminant function analysis.24 Continuous variables and binary variables remained untransformed, but
categorical variables were transformed into sets of binary variables
prior to inclusion. This resulted in 796 variables for analysis. A stepwise discriminant function analysis was conducted using SPSSv15 to
determine the variables that maximally separated the 10 groups in
the consensus grouping. For purposes of interpretation, the result-
Figure 2.
Consensus groupings of segments on a latitude-longitude plot and transposed onto a road network map
Table 1.
Correlations of IMI Items With Discriminant Functions
Variable Name
Mobile/manufactured home
Residences for seniors
Single-family detached houses
Single-family duplex
High school
Med/heavy industry
Lake
White painted lines
Blank walls
On-street parking
Speed bump/hump/dip*
Interesting architecture
Easy for walking
Road markers
No curb cuts
Pedestrian-activated signal
Safe to cross
Convenient to cross
Decorative sidewalk
Sidewalk in good condition
Narrow sidewalk
Shade from trees
Landscaping
Highway
Figure 4 Label
MobileHome
SeniorRes
SingleFam
Duplex
HighSchool
Industry
Lake
PaintedLines
BlankWalls
OnStrPk
TrafficCalm
Arch
EasyWalk
Markers
NoCurbCuts
PedActSignal
SafetoCross
ConvenCross
SWDec
SWGood
SWDNarrow
TreeShade
Landscaping
Highway
1
0.05
0.10
0.80
0.12
0.07
-0.84
0.18
0.10
-0.23
0.92
-0.12
0.65
0.64
0.01
-0.27
0.02
0.36
0.38
-0.06
0.33
0.46
0.47
0.60
-0.24
2
-0.05
0.09
0.29
-0.06
0.02
0.17
0.14
0.01
-0.03
-0.07
0.03
0.33
-0.19
-0.09
0.42
-0.65
0.55
-0.09
-0.72
-0.29
0.17
0.16
0.56
0.05
3
0.73
-0.08
-0.09
0.30
-0.15
0.03
-0.20
0.27
-0.03
-0.02
0.00
-0.25
-0.22
-0.13
0.46
-0.16
-0.03
-0.35
-0.10
-0.08
-0.11
-0.23
-0.19
0.02
Discriminant Function
4
5
6
0.17
-0.01
-0.05
0.21
0.11
-0.64
0.03
-0.09
0.10
0.14
0.06
-0.09
0.29
0.22
0.32
-0.05
-0.01
-0.04
0.02
0.54
0.08
-0.08
-0.05
0.05
0.34
0.10
0.30
0.04
-0.04
0.01
-0.02
-0.01
-0.02
-0.28
0.39
-0.01
-0.05
-0.26
-0.07
0.41
0.26
0.33
-0.32
0.27
-0.11
-0.01
0.02
-0.15
-0.10
-0.21
0.01
0.16
-0.27
0.07
-0.13
0.11
-0.13
-0.48
-0.16
0.35
0.04
-0.54
0.10
0.37
-0.31
-0.01
-0.13
0.19
0.00
-0.01
0.02
0.00
7
0.28
-0.42
-0.08
-0.02
0.09
0.06
0.57
-0.04
-0.21
0.04
0.04
0.08
-0.01
-0.14
0.03
0.02
0.29
0.17
0.03
-0.31
-0.24
-0.17
-0.13
-0.02
8
0.12
0.19
-0.13
0.20
0.36
0.06
-0.12
0.09
-0.31
0.08
0.06
0.02
0.15
-0.02
-0.17
0.02
-0.01
0.40
0.05
0.38
0.07
0.14
0.22
-0.02
9
-0.24
-0.18
0.02
0.05
0.02
-0.02
0.04
0.07
-0.18
0.11
0.01
-0.13
-0.21
-0.16
0.51
0.02
-0.14
-0.23
0.07
-0.14
0.07
0.17
0.09
-0.01
* This variable does not appear in the figures.
ing discriminant functions were calculated, and correlations between
them and IMI variables as well as the vectors specifying each of the
10 groupings were calculated. As an aid to interpretation, biplots
(two-dimensional graphs) of these correlations containing markers
for both groups and variables were examined.25
Phase four: Community Working Group (CWG)
workshops
We believe the methods detailed above provided a meaningful
reduction of the information from the IMI assessment of
Bonnyville into a form that could be effectively communicated to
community partners, especially in this community, which did not
have administratively defined neighbourhoods. We now wished to:
1) determine whether the partners would find these results compelling and valid; 2) discover how our partners would characterize
the groupings and whether this would provide additional information about the groupings not captured by the IMI (i.e., based on
the expert knowledge of people local to the community); and
3) elucidate whether the information would prove useful in planning health promotion interventions with the community.
At a regular meeting of the Bonnyville CWG (CHBE’s community partners) in spring of 2009, a workshop was held to discuss the
results of the IMI assessment. We began by requesting that the
CWG members each conduct their own grouping task using just
their own local knowledge and then describe the groupings that
emerged. A preliminary facilitated discussion with CWG members
identified some types of information that might be considered
when dividing the community into groupings. The CWG was then
divided into two separate groups of four members each to independently create their groupings on an area map provided by the
research team.
RESULTS
Analytic groupings
Figure 2 presents the 10 consensus groups from the cluster analysis of the individual ratings and, to convey perspective, also portrays the groupings on a road map.
The stepwise discriminant function analysis provided 24 statistically significant variables to separate the 10 groupings. All of the
Table 2.
Correlations of Group Membership Vectors With Discriminant Function
Group
1
2
3
4
5
6
7
8
9
10
1
-0.87
-0.03
-0.25
0.17
0.16
0.23
0.27
0.17
0.20
0.14
2
0.16
-0.85
0.00
-0.02
0.04
0.19
0.12
0.13
0.07
-0.08
3
0.00
-0.09
0.10
-0.21
-0.22
-0.11
0.13
-0.09
-0.04
0.71
4
-0.09
-0.10
0.16
0.26
0.39
-0.21
-0.27
0.21
-0.42
0.16
Discriminant Function
5
-0.05
0.09
0.20
-0.53
0.26
0.36
-0.10
0.10
-0.08
0.00
9 possible discriminant functions were statistically significant, and
the canonical correlations ranged from 0.96 to 0.44. These discriminant functions correctly classified 80.7% of segments to the
consensus groupings. Table 1 presents the correlations between the
discriminant functions and the 24 variables included in the stepwise analysis.
Table 2 presents the correlations between the group vectors and
discriminant functions for the 10 groups in the consensus groupings.
To demonstrate the interpretation of these tables, we note from
Table 1 that the variables “Single family residences”, “On-street
parking”, “Interesting architecture”, “Easy for walking” and “Landscaping” are highly positively correlated and that the presence of
“Medium/heavy industry” is highly negatively correlated with
Function 1. Table 2 shows a strong negative correlation between
Function 1 and Group 1. This suggests a higher proportion of
industrial land uses in Group 1 and a relative absence of singlefamily homes, on-street parking, interesting architecture, easy walking and attractive landscaping. Making interpretations of this type
can be greatly simplified by examining pairwise graphs such as
Figure 3, which plots the variable and group correlations from
Tables 1 and 2 for Functions 1 and 2 on the same graph.
Where variables and groupings are located close to each other
on a dimension, this can be interpreted as showing that the variable is characteristic of the grouping. Alternatively, variables located at a greater distance from a group demonstrate that the grouping
is not associated with having high values on the given variable.
We briefly summarize interpretations of the remaining functions
below. Function 2 differentiates Group 2 from the other groups by
the presence of fair sidewalks and pedestrian-activated signals.
Function 3 differentiates Group 10 from the other groups by having mobile homes. Function 4 differentiates Groups 6, 7 and 9 from
Groups 4, 5 and 8 by having attractive architecture, good sidewalks
and not having curb cuts, and by the absence of tree shade, convenient intersection crossings, seniors’ residences, high schools and
blank walls. Function 5 differentiates Group 4 from Groups 3, 6
and 8 on the basis of convenient intersection crossings, tree shade,
easy walking and narrow sidewalks, and the absence of curb cuts,
nice architecture and lake proximity. Function 6 differentiates
between Group 8 and Groups 3 and 5 according to the presence of
seniors’ residences and the absence of good sidewalks, blank walls
and neighbourhood markers. Function 7 contrasts lake proximity
with proximity to seniors’ residences and thereby separates Groups
4 and 6 from Group 8. Function 8 distinguishes Group 5 from the
other groups by the presence of high school(s), good sidewalks and
convenient crossings (most completely from Groups 3, 4 and 6, in
which these features are absent). Finally, Function 9 distinguishes
between Groups 7 and 9 primarily on the basis of the presence of
curb cuts.
6
-0.08
-0.08
0.19
0.00
0.34
-0.14
0.08
-0.51
0.16
-0.03
7
0.11
0.01
-0.30
0.13
0.07
0.32
-0.19
-0.24
-0.12
0.14
8
0.14
0.02
-0.36
-0.20
0.26
-0.23
0.06
0.10
0.11
0.05
9
0.01
0.02
-0.05
-0.03
0.01
0.01
0.34
-0.07
-0.31
-0.08
CWG groupings
Overall, there was considerable agreement between the groupings
and descriptions generated by the CWG and those generated by
our analytic methods. Elements that were common included: gridstyle development (e.g., well-connected sidewalks, shorter routes);
spaghetti roadways (e.g., cul-de-sacs, longer routes); beautification
(e.g., street flowers, bricked sidewalks, sidewalk lighting); pedestrianfriendly design (e.g., marked crosswalks, curb cuts); and sidewalk
presence. Descriptors that were not directly captured through the
IMI assessment but that the CWG considered important included
characteristics of: traffic (e.g., volume and speed); wheelchair
friendliness; locally owned establishments; historic references
(e.g., development eras); population characteristics (e.g., age, socioeconomic features, friendliness); housing styles (e.g., unique heritage houses, bland subdivision houses); and future community
developments.
We then presented the results of our analytic methods to the
CWG for its feedback. The similarity between the results of the
CWG analysis and of the analytic methods was noted immediately. However, specific characteristics of our analytic methods were
also praised. For example, the CWG was quick to validate the distinctions between Group 10 and Group 7, which neither CWG subgroup had noted in its own groupings.
DISCUSSION
Knowledge exchange between the analytic team and CWG was
important for dissemination of research results, but it also provided opportunity for the validation of those results. On one hand,
the groupings, initially created according to their composition of BE
features by the research team (an “outsider” view), were demonstrated to have community relevance. On the other hand, contextual information provided by partners (an “insider” view)
considerably enhanced the interpretation of the groupings for the
research team. The insider view provided details about the BE that
could not be readily, if at all, observed from the street level, suggesting an inherent flaw in observational studies of this kind. We
argue that modification of tools like the IMI should explicitly
include a process by which communities can add their own descriptors or a process that at least explores other insider sources of information. This would increase utility of the results for the
communities, and for those interested in participatory research
would help to solidify relationships with partners.
The process of information sharing and discussion was critical
to identifying the priority for community intervention. Near the
end of the workshop, a CWG member noted that one community
grouping contained a cluster of residences in which seniors lived.
It was also noted that this and the adjacent groupings had a lowquality sidewalk network (e.g., poor condition; low sidewalk pres-
Figure 3.
Biplot of discriminant Functions 1 and 2*
* Variables with small factor loadings are not shown.
ence; poorly connected sidewalks), and, as a result, seniors living in
that “neighbourhood” would experience difficulty accessing various community destinations necessary for their daily lives. From
this discussion, CWG members formed the concept of “you can’t
get there from here” to encapsulate the idea that if the area where
you live is not connected to destinations via a sidewalk network of
high(er) quality, the likelihood that you would choose to walk to
those destinations would be decreased. The CWG decided to have
a map created to show high-quality walking routes and to install
benches at key locations on these routes to allow seniors to rest
while in active transit. Members of the CHBE team have since
developed that map and undertaken installation of benches in partnership with the CWG.
Overall, we believe the analytic methods used to summarize the
BE assessment were sufficiently accurate and compelling to provide
a useful context for a dialogue between the research team and the
CWG. This dialogue allowed a bidirectional flow of information
that enhanced understanding on both sides and also directly facilitated action by contributing to the creation of an intervention
with the potential to enhance the health and well-being of the
community.18,19 Creating neighbourhood groupings based on BE
features is particularly useful for smaller or semi-rural communities that do not employ traditional administrative neighbourhood
boundaries and for larger communities that want to define neighbourhoods in a geographically meaningful way when developing
and implementing health promotion activities.
To our knowledge, we are the first research group to explore a
method for reducing data related to micro-features of the BE for
the purposes of health promotion intervention and creating dialogue with community stakeholders. The approach extends traditional geographic analyses of BE data26-29 that explore relations
between health and place. This paper contributes analytic and participatory techniques that can be paired or used independently to
communicate how non-administratively defined areas vary. These
BE neighbourhoods can be compared to determine the potential
each area has for interventions to support health. From this, the
type and location of interventions can be prioritized and stakeholder engagement fostered.30
Much remains to be done. It is unclear whether the discriminant
functions derived by our analytic methods reflect general BE char-
acteristics or whether they are specific to semi-rural communities in
Alberta or even to the town of Bonnyville. Thus, future work will
extend these methods to other Alberta communities. Systematic
observation exercises can be costly and time-intensive, especially if
a community does not have the resources for extensive training or
the analytic capacity to deal with the resultant dataset. Thus,
exploring the relative importance of individual variables across
communities will inform the development of observation tools that
collect a reduced number of variables. A condensed tool would
greatly increase the ability of a community to collect these data
outside of a research partnership. We also intend to formalize the
community workshop process to obtain useful quantitative data to
further examine the validity of groupings. In the current research
we did not have access to individual health information from residents in order to determine whether living within a particular area
has implications for health or healthy activity. We intend to examine this question in future research as well.
CONCLUSIONS
This project has demonstrated that it is possible to use principled
quantitative methods to reduce large amounts of BE information,
collected using inventories such as the IMI, into meaningful summaries. These summaries, or BE neighbourhoods, are inherently
valuable for initiatives bridging municipal planning and community health. Ideally, they can be enhanced and contextualized by
local knowledge provided by community stakeholders through
methods of participatory research. We have also demonstrated that
the overall research process can catalyze discussion among community stakeholders for the purposes of developing interventions
into the built environment to promote health at the community
level.
REFERENCES
1.
World Health Organization. Global Status Report on Noncommunicable Diseases 2010. Description of the Global Burden of NCDs, Their Risk Factors and
Determinants, 2011. Available at: http://www.who.int/nmh/publications/
ncd_report2010/en/index.html (Accessed December 21, 2011).
2. Waters E, de Silva-Sanigorski A, Hall BJ, Brown T, Campbell KJ, Gao Y, et al.
Interventions for preventing obesity in children. Cochrane Database Syst
Rev 2011;12.
3. Wall M, Hayes R, Moore D, Petticrew M, Clow A, Schmidt E, et al. Evaluation
of community level interventions to address social and structural determinants of health: A cluster randomised controlled trial. BMC Public
Health 2009;28(9):207-17.
4. Raine K, Spence JC, Church J, Boulé N, Slater L, Marko J, et al. State of the Evidence Review on Urban Health and Healthy Weights. Ottawa, ON: Canadian
Institute for Health Information, 2008.
5. Egger G, Swinburn B. An ecological approach to the obesity pandemic.
BMJ 1997;315:477-80.
6. Swinburn B, Egger G. Preventive strategies against weight gain and obesity.
Obes Rev 2002;3:289-301.
7. Booth KM, Pinkston MM, Carlos Poston WS. Obesity and the built environment. J Am Diet Assoc 2005;105(5 Suppl 1):S110-17.
8. Cummins S, Macintyre S. Food environments and obesity – Neighborhood or
nation? Int J Epidemiol 2006;35(1):100-4.
9. Kumanyika SK, Obarzanek E, Stettler N, Bell R, Field AE, Fortmann SP, et al.
Population-based prevention of obesity: The need for comprehensive promotion of healthful eating, physical activity, and energy balance. A scientific statement from American Heart Association Council on Epidemiology and
Prevention, Interdisciplinary Committee for Prevention (formerly the expert
panel on population and prevention science). Circulation 2008;118(4):42864.
10. Duncan MJ, Spence JC, Memmery WK. Perceived environment and physical
activity: A meta-analysis of selected environmental characteristics. Int J Behav
Nutr Phys Act 2005;2:11-20.
11. Spence JC, Plotnikoff RC, Rovniak LS, Martin Ginis KA, Rodgers W, Lear SA.
Perceived neighbourhood correlates of walking among participants visiting
the Canada On the Move website. Can J Public Health 2006;97(Suppl 1):S36-40.
12. Frank LD, Schmid TL, Sallis JF, Chapman J, Saelens BE. Linking objectively
measured physical activity with objectively measured urban form: Findings
from SMARTRAQ. Am J Prev Med 2005;28(2 Suppl 2):117-25.
13. Hoehner CM, Brennan Ramirez LK, Elliott MB, Handy SL, Brownson RC.
Perceived and objective environmental measures and physical activity among
urban adults. Am J Prev Med 2005;8(2 Suppl 2):105-16.
14. The Ottawa Charter for Health Promotion. Ottawa: Health and Welfare
Canada, 1986. Available at: http://www.who.int/healthpromotion/conferences/
previous/ottawa/en/print.html (Accessed December 21, 2011).
15. Brownson RC, Hoehner CM, Day K, Forsyth A, Sallis JF. Measuring the built
environment for physical activity: State of the science. Am J Prev Med
2009;36(Suppl 4):S99-123.
16. Lytle LA. Measuring the food environment: State of the science. Am J Prev
Med 2009;36(Suppl 4):S134-44.
17. Day K, Boarnet M, Alfonzo M, Forsyth A. The Irvine-Minnesota Inventory to
measure built environments: Development. Am J Prev Med 2006;30(2):14452.
18. Nykiforuk CIJ, Schopflocher D, Vallianatos H, Spence JC, Raine KD,
Plotnikoff RC, et al. Community Health and the Built Environment: Examining place in a Canadian chronic disease prevention project. Health Promot
Int 2012; first published online January 6, 2012 doi:10.1093/heapro/dar093.
19. Israel BA, Schulz AJ, Parker EA, Becker AB. Review of community-based
research: Assessing partnership approaches to improve public health.
Annu Rev Public Health 1998;19:173-202.
20. Moffatt L. Chronic Disease Prevention Alliance of Canada. Knowledge
Exchange in Health Promotion: Theoretical Models and Examples. Ottawa,
ON: Chronic Disease Prevention Alliance of Canada, 2007. Available at:
http://www.cdpac.ca/media.php?mid=274 (Accessed December 21, 2011).
21. Town of Bonnyville. Bonnyville in Profile. Available at: http://www.town.bonnyville.ab.ca/index.php/living-in-bonnyville/about-bonnyville/bonnyvillein-profile (Accessed December 21, 2011).
22. Rosenberg S, Kim MK. The method of sorting as a data-gathering procedure
in multivariate research. Multivariate Behav Res 1975;10(4):489-502.
23. Ward JH. Hierarchical grouping to optimize an objective function. J Am Stat
Assoc 1963;580:236-44.
24. Izenman AJ. Modern Multivariate Statistical Techniques: Regression, Classification, and Manifold Learning. New York: Springer, 2008.
25. ter Braak CJF. Interpreting canonical correlation analysis through biplots of
structure correlations and weights. Psychometrika 1990;55(3):519-31.
26. Boone-Heinonen J, Popkin BM, Song Y, Gordon-Larsen P. What neighborhood area captures built environment features related to adolescent physical
activity? Health Place 2010;16(6):1280-86.
27. Oliver L, Schuurman N, Hall A, Hayes M. Assessing the influence of the built
environment on physical activity for utility and recreation in suburban Metro
Vancouver. BMC Public Health 2011;11(1):959.
28. Hoehner CM, Handy SL, Yan Y, Blair SN, Berrigan D. Association between
neighborhood walkability, cardiorespiratory fitness and body-mass index.
Soc Sci Med 2011;73(12):1707-16.
29. Keegan TH, Hurley S, Goldberg D, Nelson DO, Reynolds P, Bernstein L, et al.
The association between neighborhood characteristics and body size and
physical activity in the California Teachers Study cohort. Am J Public Health
[published online ahead of print August 18, 2011].
30. McCreedy M, Leslie JG. Get Active Orlando: Changing the built environment
to increase physical activity. Am J Prev Med 2009;37(6 Suppl 2):S395-402.
RÉSUMÉ
Objectifs : Les évaluations approfondies du milieu bâti résistent souvent
aux tentatives de réduction et de synthèse des données. Nous avons
cherché à élaborer une méthode de réduction des données sur le milieu
bâti qui permette de communiquer efficacement ces données aux
chercheurs et aux acteurs locaux. Notre objectif est de faire comprendre
comment on peut utiliser ces données pour créer des regroupements de
quartiers fondés sur les caractéristiques du milieu bâti, et que le processus
de discussion des quartiers avec les acteurs locaux peut entraîner la mise
au point d’interventions de promotion de la santé renforcées par un
apport communautaire.
Méthode : À l’aide de la liste de critères Irvine-Minnesota Inventory
(IMI), nous avons évalué 296 segments d’une communauté semi-rurale
de l’Alberta. Des évaluateurs experts ont « créé » des quartiers en
examinant les données. Ensuite, nous avons élaboré un regroupement
consensuel au moyen d’une analyse en grappes, et réduit le nombre de
variables IMI caractérisant les quartiers au moyen d’une analyse
discriminante multiple.
Résultats : Les 296 segments ont été réduits par consensus à un
ensemble de 10 quartiers, lesquels se distinguent les uns des autres selon
9 fonctions construites à partir de 24 variables IMI. Des biplots de ces
fonctions ont été un moyen efficace de résumer et de présenter les
résultats de l’évaluation communautaire, et ont stimulé l’action
communautaire.
Conclusions : Il est possible d’utiliser des méthodes quantitatives
raisonnées pour réduire de grandes quantités d’information sur le milieu
bâti en résumés signifiants. Ces résumés, ou « quartiers selon le milieu
bâti », ont été utiles pour catalyser des actions avec les acteurs locaux et
ont mené à l’élaboration d’interventions sur le milieu bâti favorisant la
santé.
Mots clés : milieu bâti; méthodes quantitatives; promotion de la santé;
échange des connaissances; réduction de l’obésité
Examining Aspects of the Built Environment: An Evaluation of a
Community Walking Map Project
Candace I.J. Nykiforuk, PhD,1 Laura M. Nieuwendyk, MSc,1 Shaesta Mitha, MPH,2 Ian Hosler3
ABSTRACT
Objective: Interventions that address the built environment present an opportunity to affect behaviours such as physical activity. The purpose of this
study was to evaluate a community walking map developed for eight neighbourhoods in the City of Edmonton (COE).
Method: A walking map developed in partnership with the COE’s Walkable Initiative was distributed to 11,994 households across eight
neighbourhoods in July 2010. In total, 149 respondents completed an online follow-up survey that assessed the effectiveness of the walking maps in
influencing physical activity.
Results: Of the 149 respondents, 89 (59.7%) reported that they had received a copy of the map, and 60 (40.2%) reported that they had not. Of those
who had a copy, 76.4% (n=68) indicated that the routes and destinations on the map encouraged them to walk more in the community, 64.0% (n=57)
stated they would walk more often to get to destinations, and 55.1% (n=49) indicated they would walk more often for physical activity or exercise as a
result of having a copy of the map. Finally, 91.0% (n=81) stated that they found the map to be useful, as it provided walking routes (60/81, 74.1%,)
and places to go in the community (57/81, 70.4%). Of those who did not receive a copy, 95.0% (n=57) indicated that they would use a community
walking map.
Conclusion: This evaluation demonstrated that a community walking map was a valuable tool for not only encouraging walking for physical activity
but also motivating individuals to explore their communities and visit local community destinations.
Key words: Built environment; walking; physical activity; health promotion; urban health
C
oncern about obesity rates in Canada has resulted in various
community-based health promotion interventions to
increase physical activity.1-3 Despite growing evidence of the
benefits of physical activity in preventing and controlling chronic
diseases,3 most Canadian adults are not sufficiently active to reap
the health benefits of a physically active lifestyle.4,5
Walking is the most frequent and preferred form of physical
activity across both sexes and different ages and income levels.6,7
Walking is accessible, as it requires no special skills or equipment,
is affordable and can be made easily routine, particularly if done
for active transportation.8 Thus, the quality of walking routes and
the presence, type and convenience of destinations in a community
affect how much people walk.6 Clearly, identification of safe and
convenient walking routes that enable residents to reach destinations is crucial to promote walking for physical activity.
Community-based interventions that target walking for recreation or active transportation support increased physical activity.9
They engage stakeholders and are tailored to consider community
characteristics and needs with the goal of reducing the population’s
risk of disease.10 Walking trails are useful community-based physical activity interventions, but despite the apparent beneficial
effects, may be under-used once implemented.11
In the current project, community consultation revealed that
development of a map of local walking routes was more viable and
economically feasible than developing walking trails. This paper
will discuss the evaluation of a walking map developed for one geo-
graphic community in the City of Edmonton (COE), Alberta. The
evaluation sought to identify whether the map encouraged walking
among residents and to assess its value as a tool for informing them
about community assets and destinations.
METHODS
Background
Researchers collaborated with the COE’s Walkable Edmonton initiative for the evaluation. Community walking maps provide routes
throughout the designated area while guiding residents to specific
destinations or points of interest. The maps produced as part of the
“Communities on Foot” Walking Map Series (http://www.edmonton.ca/community-walking-maps.aspx) aim to: 1) encourage citizens
of all ages to partake in active transportation, particularly walking;
2) encourage community members to walk and explore their neighAuthor Affiliations
Edmonton, AB
2. Health Facility Planning, Regional Municipality of Peel, Mississauga, ON
3. Walkable Edmonton, City of Edmonton Community Services, Edmonton, AB
Correspondence: Candace I.J. Nykiforuk, Asst. Professor, Centre for Health
Promotion Studies, School of Public Health, University of Alberta, 3-258 ECHA, 1140587 Ave., Edmonton, AB T6G 1C9, Tel: 780-492-4109, Fax: 780-492-0364,
E-mail: [email protected]
Acknowledgements: Funding for this project was provided by the Heart and Stroke
Foundation of Canada in partnership with the Canadian Institutes of Health Research.
I. Hosler is a community partner who was employed by the walking map initiative.
COMMUNITY WALKING MAP EVALUATION
Table 1.
Categorized Questions
Focus Area for Evaluation
Awareness
• Did respondents have a copy of the walking map?
• Were respondents aware that a map had been developed for their
neighbourhood?
Purpose
To determine the number of individuals who had a copy of the map and whether
they were aware that the map had been produced.
Map distribution
• How did people get their copy of the map?
• How would they like to get a copy of the map in the future?
To determine what tactics were effective in the dissemination of the map and
how residents would like to receive it.
Short-term change
• Did the map influence behaviour in the short term?
• Did respondents walk more to local destinations or for recreation?
• Did the map actually get respondents walking (distinction between
exercise and purposeful walking)?
To compare the amount of time the respondents had the map with the number
of routes they tried and to determine whether people actually walked the routes.
Long-term change
• Did the map influence behaviour change in the long term?
To determine the long-term behaviour change produced by the map.
Personal impact
• Did respondents find this product a useful resource?
To determine whether respondents found the map to be a useful tool in
identifying safe walking routes in their community.
Community impact
• Is this a worthwhile project or resource for the community?
• What other benefits for the community were associated with doing
this project?
• Have respondents visited local destinations highlighted in the map?
To determine the value respondents placed on the map and whether they had
visited local destinations shown on the map.
Map perception
• What aspects of the map did the respondents like or dislike?
To determine what aspects of the map were viewed as valuable by the
respondents and what they did not like.
Map improvement
• How could this be a more useful tool/resource?
A list of potential map improvements was provided to respondents to check all
that apply in terms of the types of change that would make the map a more
useful tool.
Demographic information
• Demographic information
Table 2.
General demographic data were necessary to obtain information on the types of
individual who responded.
Profile of Survey Respondents
Respondent Characteristics
Age
Household income (N=149)
Neighbourhood (N=149)
Male
Female
Prefer not to answer
Total
18-34
35-54
55+
Total
Less than $23,000
$24,000-$60,000
More than $60,000
Don’t know
Missing
Total
From walk map area
Outside walk map area
Don’t know
Missing
Total
Total
n (%)
43 (28.9)
103 (69.1)
3 (2.0)
149 (100.0)
34 (22.8)
75 (50.3)
36 (24.2)
4 (2.7)
149 (100.0)
14 (9.4)
51 (34.2)
51 (34.2)
1 (0.7)
31 (20.8)
1 (0.7)
149 (100.0)
99 (66.4)
37 (24.8)
4 (2.7)
9 (6.0)
149 (100.0)
bourhoods, parks, trails and business districts; and 3) foster community engagement in building a walkable city.12 The walking maps
are created by residents for residents: residents are recruited to participate in map production and are given the task of identifying
key walking routes and destinations within their community.
Setting
The evaluation focused on a walking map recently prepared for a
community comprising eight contiguous neighbourhoods in innercity Edmonton. This community is proximal to the downtown core
and contains some of the city’s oldest neighbourhoods. The area
includes a range of housing styles, from small bungalows to larger
two-storey homes, and a variety of business districts. This commuS68 REVUE CANADIENNE DE SANTÉ PUBLIQUE • VOL. 103, SUPPLÉMENT 3
Walking Map
(YES) n (%)
31 (34.8)
57 (64.0)
1 (1.1)
89 (100.0)
22 (24.7)
48 (53.9)
18 (20.2)
1 (1.1)
89 (100.0)
7 (7.9)
34 (38.2)
28 (31.5)
1 (1.1)
18 (20.2)
1 (1.1)
89 (100.0)
68 (76.3)
18 (20.2)
1 (1.1)
2 (2.2)
89 (100.0)
Walking Map
(NO) n (%)
12 (20.0)
46 (76.7)
2 (3.3)
60 (100.0)
12 (20.0)
27 (45.0)
18 (30.0)
3 (5.0)
60 (100.0)
7 (11.7)
17 (28.3)
23 (38.3)
–
13 (21.7)
–
60 (100.0)
31 (51.6)
19 (31.6)
3 (5.0)
7 (11.7)
60 (100.0)
nity has a diverse population with average incomes significantly
lower (by about 35%) than the total average income of the municipality.13 The community walking map depicts 10 sample walking
routes ranging in length from 1.4 to 5.0 km with at least one route
through each neighbourhood. Historical information about each
neighbourhood is provided, and key community locations and
amenities are indicated, including community centres, schools,
health centres, libraries, bus and light rail transit stations and stops,
grocery stores, bakeries, hotels and recreational facilities. All information is superimposed onto an aerial photograph of the area,
which also shows building footprints and road layouts (map can be
viewed at http://www.edmonton.ca/transportation/AvenueCommunitiesWalkingMap.pdf).
Table 3.
Summary of Survey Responses From Those
Respondents Who Had a Map
Question
Total n (%)
How did you get your copy of your community walking map?
In the mail
54 (60.7)
From a friend or family member
5 (5.6)
From a place in my community (e.g., library)
16 (18.0)
Other (e.g., internet)
10 (11.2)
Missing
4 (4.5)
Total
89 (100.0)
Have you tried any of the walking routes identified?
Yes
43 (48.3)
No
46 (51.7)
Total
89 (100.0)
How many of the routes have you used in the maps?
1
15 (34.9)
2
10 (23.3)
3
12 (27.9)
4
1 (2.3)
5
3 (7.0)
Don’t know
2 (4.7)
Total
43 (100.0)
Have you used the maps to discover new places to visit in
your community (e.g., the library, parks or coffee shops)?
Yes
39 (43.8)
No
50 (56.2)
Total
89 (100.0)
How many of the community destinations have you visited
since getting the map?
1 to 2 locations
22 (56.4)
3 to 4 locations
10 (25.6)
5 to 6 locations
3 (7.7)
6 to 10 locations
2 (5.1)
Don’t know
2 (5.1)
Total
39 (100.0)
Have the routes and destinations on the map encouraged
you to walk more in your community?
Yes
68 (76.4)
No
21 (23.6)
Total
89 (100.0)
How have the maps encouraged you to walk more?
(check all that apply)
To visit places in my community (e.g., library or coffee shop) 34 (50.0)
To get more exercise
45 (66.2)
To get out and enjoy my community
40 (58.8)
To learn about the history of my community
22 (32.4)
To become familiarized with my community
45 (66.2)
To get to know other members in my community
9 (13.2)
Other
5 (7.4)
Do you think that you will walk more often to get to
destinations because you have the map?
Yes
57 (64.0)
No
31 (34.8)
Missing
1 (1.1)
Total
89 (100)
Do you think you will walk more often for physical activity
or exercise because you have the map?
Yes
49 (55.1)
No
39 (43.8)
Missing
1 (1.1)
Total
89 (100)
Do you find this map useful?
Yes
81 (91.0)
No
8 (9.0)
Total
89 (100)
Yes
Provides places to go in the community
57 (70.4)
Provides walking routes
60 (74.1)
Other (e.g., length of route is provided, useful for
visitors/new community members, good for exploring
community, used for cycling routes, and crosswalk
locations provided)
15 (18.5)
No
Does not provide places I want to go in the community 1 (12.5)
Does not provide me with appropriate walking routes
2 (25.0)
Other (e.g., unsafe areas to walk, lived in the area for
a while – don’t need a map to walk)
6 (75.0)
Design
In July 2010, the Canada Post Unaddressed Admail System, which
delivers mail through generic postal codes, was used to distribute
11,994 walking maps to all households (houses and apartments) in
the mapped area. After this mail-out, a cross-sectional, post-test-
only survey was used to collect information on the effectiveness of
the maps in encouraging walking in the community. The
10-minute survey was available online through a link from the COE
website. A variety of methods were used to recruit the area’s adult
population while retaining opportunities for minority or harderto-reach populations to participate.
Recruitment methods included a hot-link button on the COE
website, manned poster displays at key locations and distribution
of two separate reminder postcards (including information about
the map, survey and locations with free internet access) to all
households. To gather feedback from residents who did not receive
a copy of the map, student volunteers were present at key locations
in the community (e.g., library, grocery store, ethnic centre) during
the month of October 2010 to hand out maps and postcards to
encourage survey participation. Given the low-income status of the
neighbourhood, internet access to complete the survey was of significant concern. A toll-free number was established so that those
without internet access could complete the survey over the telephone with a member of the research team. Discussion with community partners revealed that incentives should be provided to
encourage residents to complete the survey. To meet this need, an
early bird draw prize ($100 gift certificate to a local grocery store)
and grand prize (mountain bike and helmet valued at $500) were
provided.
Ethical clearance for the project was received from the Health
Research Ethics Board (Panel B), University of Alberta.
Sample
Respondents were recruited from the eight contiguous neighbourhoods represented on the map and from surrounding neighbourhoods. A total of 155 people, aged 18 years or older, participated
(i.e., 1% of households in the map area).
Measures
Survey questions were developed in collaboration with Walkable
Edmonton and other community partners involved in the map
production to ensure that the findings would be relevant for program providers. New questions were developed to meet community needs, as a literature review revealed no previous indicators
reported from similar evaluations. Survey questions were simply
stated and attempted to measure awareness of the map, participation in map development, short-term behaviour change, community impact and suggestions for map improvements. A separate set
of questions was developed for respondents who did not receive a
copy of the map. Table 1 summarizes the survey questions.
Respondents were asked to identify which neighbourhood they
resided in and whether they had a copy of the map. While the
maps were delivered to every household in the community, respondents might not have received one or might have discarded it.
Respondents who did not have a copy of the map were directed to
a shorter version of the survey that asked whether they would find
the map useful, had ideas on the best way to share it with the community, and would like to receive a copy of it; those who did were
directed to Walkable Edmonton. Respondent demographic information (age, household income and number of individuals residing in the household) was collected in both versions to facilitate
characterization of respondents and groups not reached through
the evaluation design.
Table 4.
Summary of Survey Responses From Those Respondents Who Did Not Have a Map
Question
Were you aware that a community walking map was created for your community with routes and key destinations?
Yes
No
Total
Would you like to have a map of your community with walking routes and interesting destinations (e.g., coffee shops and parks) identified?
Yes
What would be the best way for us to get the community walking maps to people in the community? (check all that apply)
Through the mail
Have it available at community locations (e.g., coffee shop, library or grocery store)
E-mail
Online
Other
Would you use a community map?
Yes
No
Total
Please tell us why you would not use a community map with walking routes and destinations?
Not interested in my community
Other
Missing
Total
Analysis
Survey data were analyzed using SPSS v.18.0. Descriptive statistics
and frequencies were calculated.
RESULTS
Of the 155 respondents, 57.4% had a copy of the walking map,
38.7% did not, and 3.9% opted not to complete the survey past the
first question and were excluded from further analysis. Of the
resulting 149 respondents, there were more females (69.1%) than
males (28.9%), and a range of income brackets was represented.
The majority of respondents (66.4%) lived in the map area, and
24.8% were from surrounding neighbourhoods. Only four respondents had participated in the development of the map. Table 2 provides the demographic profile of respondents.
Table 3 summarizes survey responses for the 89 respondents who
had a walking map. Most of those respondents (60.7%) received
their maps through the mail, and 18.0% obtained a copy from community destinations. In total, 48.3% had tried a walking route identified on the map, with variation in the number of routes tried. Of
the 43.8% of respondents who used the maps to discover new
places in the community, 56.4% visited one to two locations, and
38.4% visited multiple locations.
The majority of respondents (76.4%) agreed that the routes and
destinations on the map encouraged them to walk more in the
community, their rationales including the desire to: get more exercise (66.0%); become familiarized with the community (66.2%); get
out and enjoy the community (58.8%); visit places in the community (e.g., library or coffee shop) (50.0%); and learn about the community’s history (32.4%) (Table 3). Overall, respondents stated that,
as a result of having the map, they would walk more often to get
to destinations (64.0%) and for physical activity or exercise (55.1%).
Finally, 91.0% stated that they found the map useful as it identified
walking routes (74.1%) and places to go in the community (70.4%).
Table 4 summarizes survey data for respondents who did not
have a map, of whom 56.7% were not aware that a walking map
had been created for their community. All of these respondents
expressed interest in obtaining the map. Respondents suggested
sharing the map through community locations (e.g., coffee shop,
library or grocery store) (73.5%); postal mail (70.6%); online
(52.9%); and E-mail (32.4%). Overall, 95% of these respondents
stated that they would use a community walking map.
Total n (%)
26 (43.3)
34 (56.7)
60 (100.0)
34 (56.7)
24 (70.6)
25 (73.5)
11 (32.4)
18 (52.9)
7 (20.6)
57 (95.0)
3 (5.0)
60 (100.0)
1 (33.3)
1 (33.3)
1 (33.3)
3 (100.0)
DISCUSSION
This community-research partnership was an ideal opportunity to
better understand residents’ perceptions of a community walking
map. The evaluation assessed whether the map’s routes and destinations adequately met residents’ needs and whether the map was
being utilized as intended. Further, the survey provided insight on
the reach of the map: the results describe perceptions from respondents in eligible households who did and did not receive the map.
While findings were specific to this community, a general understanding of perceptions about walking maps can help researchers
and community stakeholders develop future walking map initiatives.
Evaluation findings indicated that the postal mail-out was the
most preferred means to distribute the maps, despite initial partner
concerns that maps would be discarded (and despite the low
response rates we experienced with this approach). This finding is
congruent with similar programs, in which direct communication
was respondents’ preferred method of contact.14 Yet, as a significant
proportion of respondents did not have the map, concerted efforts
are needed to ensure that the map is prominent among the plethora of advertisements received in the mail. The map also reached people from outside the map area, likely because of its availability at
various community destinations and online. Thus, a walking map
highlighting routes, destinations and interesting community information may be a useful way to attract visitors to the community.
Approximately half of respondents who had the map attempted
one or more walking routes. Thus, short-term behaviour change
was fostered, especially among those who reported that the map’s
routes and destinations encouraged them to walk more. While the
evaluation did not assess previous walking behaviours, respondents
expressed an inclination to walk more because they now had the
map. This is consistent with previous studies’ findings that individuals given maps of walking routes were twice as likely as controls
to walk to work during a six-month follow-up.15 One study has even
indicated that short-term behaviour change related to increased
walking was consistent with results at 10-year follow-up.16 Given
this potential, future research should explore the impact of walking
interventions on long-term behaviour change.17
The participatory nature of this evaluation helped foster resident
engagement and promote community development through stake-
holder involvement in the process. The exploration of residents’
perceptions of the map revealed that community activity was
enhanced in these neighbourhoods. This walking map demonstrated the potential to encourage residents to come out into the
community, thereby creating opportunities for increased interaction. Respondents became more aware of walking for physical
activity and of safe walking routes. Thus, the map was a powerful
tool that helped respondents recognize existing opportunities by
familiarizing them with their community and facilitating interaction with other community members.18
The majority of respondents indicated that the map was useful
because of the walking routes, but only half had attempted walking a route. Respondents had received the map a few weeks before
the survey, so may not yet have had time to try the routes.
Although a small percentage found the map not useful, their
responses to open-ended survey questions indicated that this was
because of their perceptions of the community rather than of the
map itself (e.g., respondents noted that it was unsafe to walk in the
community, did not want to visit places in the community). Some
of the walking routes did intersect with undesirable areas of this
inner-city community, however, mobilizing residents to walk
through these areas could increase street presence and informal
monitoring. Pedestrian design enhancements (e.g., pedestrianactivated crossing signals) could also make walking the routes a
more pleasant and a safer form of recreation or transportation.19
Respondents with a map provided feedback on whether the map
was of interest/use to community members. Yet, about one third
of respondents did not receive the map, and more than half of
these respondents were not aware the map existed. Examination
of map dissemination strategies offered insight into how to address
this situation. Respondents indicated that postal mail was the best
way to distribute the maps, followed by having them available at
community locations and online. Thus, a combination of dissemination methods should be considered as viable means to provide
access to the walking map, including use of alternative methods
(e.g., posters or signs, community board postings) to increase the
visibility of and community resonance with the initiative.14
Respondents’ high level of interest in obtaining and using the
walking map suggests that they are open to walking. Survey findings demonstrate that residents perceive the walking map to be a
valuable information tool about options available in the community, and the depiction of various walking routes to be appealing.
Communities should be enabled to develop walking maps as a
means to foster community engagement, increase physical activity and encourage active transportation.
Use of a community-university partnership approach strengthened
this evaluation. Collaboration with key stakeholders was crucial to
obtain project acceptance and facilitate evidence-gathering that
would be meaningful for community action.14,20 Here, evaluation
findings were used by community partners in planning future initiatives.
The examination of perceptions of respondents with and without a copy of the map was advantageous as it allowed the team to
gain a broader understanding of the overall value placed on the
map by residents. Both groups valued the map as a tool to increase
their knowledge of walking routes and local community destina-
tions. Ultimately, this information is useful for public health practitioners, community leaders and government officials when forming physical activity, active transportation or community
development initiatives.
While designed to reach as many potential respondents as possible with limited resources, this evaluation was limited by its crosssectional, post-test-only survey design. The resultant data provided
only a small snapshot of information. The survey relied on selfreported data and, to fit within the grant funding period, was available online for only a limited time (3.5 months), which began
immediately after the map had been distributed. This may not have
been an adequate amount of time to assess the impacts of the map
on residents. Limited time and passive recruitment through postal
mail along with the low socio-economic status of the area were likely the primary contributors to the very low response rate to the survey. It may be that those who responded did so because of a special
interest in walking, walking maps or their neighbourhood.
Survey participation was limited to adults aged 18 and older, but
the walking map was available to people of all ages. Consequently,
a small sample and large proportion of female respondents limited
the representativeness and generalizability of the findings. Given
the short funding period for this program, a more extensive evaluation was not possible to discover resident perceptions and detect
long-term behaviour change. Despite the shortcomings of the survey design, the approach was necessary to generate feedback on
resident perceptions of the map in order to inform other COE map
developments planned for the immediate future.
Future implications for research, practice and policy
Locally, future evaluation should consider the entire series of COE
walking maps available to better understand the impacts of walking maps in different communities within the municipality over
time. The ability to undertake such a wide-scale evaluation would
require additional resources and support from government officials.
More broadly, walking map evaluations should follow a pre-post
survey design with multiple follow-ups to assess utilization as well
as short- and long-term behaviour change. Gathering data on measured physical activity levels could enhance the proposed links
between walking maps and behaviour. Future research would benefit from multiple recruitment strategies as well as integration of
qualitative methods (e.g., focus groups) to explore utilization issues
in greater detail.
There is little research examining walking maps as an intervention tool to promote physical activity and increase community
awareness. Findings from this evaluation suggest that, when developing interventions to promote walking, practitioners should
design a tool that: provides residents with a tangible item outlining
safe walking routes; identifies pedestrian supports (e.g., crosswalk
availability); demarcates key community destinations; and considers specific community needs and characteristics (e.g., multicultural destinations). This evaluation provides preliminary evidence
that walking maps are valued by residents and are perceived as an
effective means to increase local walking.
Community-based initiatives like walking maps must be supported by healthy public policy. For example, future health promotion interventions should consider urban planning or
transportation policies as a means to support the development of
walking-friendly environments. Finally, collaboration with profesCANADIAN JOURNAL OF PUBLIC HEALTH • NOVEMBER/DECEMBER 2012 S71
sionals from various disciplines and involvement of key community partners (including residents) are also essential to map development and successful walking initiatives.
CONCLUSION
Walking is a form of physical activity accessible to individuals of all
ages and in all types of built environments. Community walking
maps are a valuable intervention to foster change in walking behaviours by informing individuals about walking route options and
motivating them to explore their communities. By recognizing the
potential of community walking maps, policy-makers and practitioners can work towards implementing this intervention to
enhance citizen engagement and promote walking for recreation
and transportation in their communities.
REFERENCES
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
Burke NM, Chomitz VR, Rioles NA, Winslow SP, Brukilacchio LB, Baker JC.
The path to active living: Physical activity through community design in
Somerville, Massachusetts. Am J Prev Med 2009;37(6 Suppl 2):S386-94.
van Sluijs EM, McMinn AM, Griffin SJ. Effectiveness of interventions to promote physical activity in children and adolescents: Systematic review of controlled trials. BMJ 2007;335(7622):703.
Warburton DE, Nicol CW, Bredin SS. Health benefits of physical activity: The
evidence. CMAJ 2006;174(6):801-9.
Vanasse A, Demers M, Hemiari A, Courteau J. Obesity in Canada: Where and
how many? Int J Obes (Lond) 2006;30(4):677-83.
Katzmarzyk PT, Gledhill N, Shephard RJ. The economic burden of physical
inactivity in Canada. CMAJ 2000;163(11):1435-40.
Moudon AV, Lee C, Cheadle AD, Garvin C, Rd DB, Schmid TL, Weathers RD.
Attributes of environments supporting walking. Am J Health Promot
2007;21(5):448-59.
Hooker SP, Cirill LA, Wicks L. Walkable neighbourhoods for seniors: The
Alameda County experience. J Appl Gerontol 2006;26(2):157-81.
Dannenberg AL, Cramer TW, Gibson CJ. Assessing the walkability of the
workplace: A new audit tool. Am J Health Promot 2005;20(1):39-44.
Moudon AV, Lee C. Walking and bicycling: An evaluation of environmental
audit instruments. Am J Health Promot 2003;18(1):21-37.
McLeroy KR, Norton BL, Kegler MC, Burdine JN, Sumaya CV. Communitybased interventions. Am J Public Health 2003;93(4):529-33.
Brownson RC, Baker EA, Boyd RL, Caito NM, Duggan K, Housemann RA,
et al. A community-based approach to promoting walking in rural areas.
Am J Prev Med 2004;27(1):28-34.
Walkable Edmonton. First Steps: Walkable Edmonton Committee 2008 Annual Report, 2008. Available at: http://www.edmonton.ca/for_residents/WalkableEdmontonAnnualReport2008.pdf (Accessed October 8, 2010).
City of Edmonton. Neighbourhood Profiles, 2010. Available at:
http://www.edmonton.ca/for_residents/neighbourhoods/ (Accessed October
25, 2010).
Cooper C. Successfully changing individual travel behavior: Applying
community-based social marketing to travel choice. J Transportation Research
Board 2007;2021(11):89-99.
15. Mutrie N, Carney C, Blamey A, Crawford F, Aitchison T, Whitelaw A. “Walk
in to work out”: A randomised controlled trial of a self help intervention to
promote active commuting. J Epidemiol Community Health 2002;56(6):407-12.
16. Pereira MA, Kriska AM, Day RD, Cauley JA, LaPorte RE, Kuller LH. A randomized walking trail in postmenopausal women: Effects on physical activity and health 10 years later. Arch Intern Med 1998;158:1695-701.
17. Williams DM, Matthews CE, Rutt C, Napolitano MA, Marcus BH. Interventions to increase walking behavior. Med Sci Sports Exerc 2008;40(7 Suppl):S56773.
18. Miller EK, Scofield JL. Slavic Village: Incorporating active living into community development through partnerships. Am J Prev Med 2009;37(6 Suppl
2):S377-85.
19. Pucher J, Dijkstra L. Promoting safe walking and cycling to improve public
health: Lessons from the Netherlands and Germany. Am J Public Health
2003;93(9):1509-16.
20. Horowitz CR, Robinson M, Seifer S. Community-based participatory research:
From the margin to the mainstream. Circulation 2009;119:2633-42.
RÉSUMÉ
Objectif : Les interventions sur le milieu bâti sont des occasions
d’influencer les comportements comme l’activité physique. Notre étude
visait à évaluer une carte de marche dans la communauté élaborée pour
huit quartiers de la ville d’Edmonton.
Méthode : Une carte de marche, élaborée en partenariat avec le projet
Walkable Initiative d’Edmonton, a été envoyée à 11 994 ménages de huit
quartiers en juillet 2010. En tout, 149 répondants ont rempli un
questionnaire de suivi en ligne évaluant l’influence de ces cartes sur leur
niveau d’activité physique.
Résultats : Sur les 149 répondants, 89 (59,7 %) ont dit avoir reçu un
exemplaire de la carte, et 60 (40,2 %) ont dit ne pas en avoir reçu. De
ceux qui en avaient un exemplaire, 76,4 % (n=68) ont indiqué que les
trajets et les points d’intérêt sur la carte les incitaient à marcher
davantage dans la communauté, 64 % (n=57) ont dit qu’ils se rendaient
plus souvent à leurs destinations à pied, et 55,1 % (n=49) ont indiqué
qu’ils marchaient plus souvent pour faire de l’activité physique ou de
l’exercice depuis qu’ils avaient la carte. Enfin, 91 % (n=81) ont dit avoir
trouvé la carte utile, car elle propose des trajets (60/81, 74,1 %) et des
endroits à visiter dans la communauté (57/81, 70,4 %). De ceux qui
n’avaient pas reçu la carte, 95 % (n=57) ont indiqué qu’ils se serviraient
d’une carte de marche dans la communauté.
Conclusion : Cette évaluation montre qu’une carte de marche dans la
communauté est un outil précieux non seulement pour encourager la
marche en tant qu’activité physique mais pour inciter les gens à explorer
leur communauté et à visiter des points d’intérêt locaux.
Mots clés : milieu bâti; marche; activité physique; promotion de la
santé; santé en zone urbaine

Print this article - Canadian Journal of Public Health

Transcription

Similar documents

Youth Asset Mapping in Toronto

Westney Heights Plaza Ajax - December 10, 2015

THS SALESAS ES SF.key

Mount Pleasant Community Brochure

The Connaught Eco-Neighbourhood Project

Being in Neighbourhood - Contemporary Practices

governance processes over the provision of local facilities for

Walker Profile - City of Edmonton

COMRADERY FROM COAST-TO

Seedling Giveaway