Classifying neighbourhoods by level of access to stores selling fresh fruit and vegetables and groceries: identifying problematic areas in the city of Gatineau, Quebec.
Gould, Adrian C. ; Apparicio, Philippe ; Cloutier, Marie-Soleil 等
The complex relationship between social and material deprivation
and health status is well documented in Canada. (1,2) Diet, and in
particular, the lack of access to an adequate diet, is just one among
the numerous factors associated with social inequalities in health. (3)
In light of the relationship between diet and health, much research has
concentrated on the different elements that affect access to food.
Disposable income is the primary factor determining whether a household
has access to a sufficient quality and quantity of food to meet the
dietary needs of its members. (4) Informational constraints also play an
important role in determining whether all members of a household have
access to a healthy diet. (5) Finally, having "physical"
access to a source of food (referred to as geographic access) has also
been the subject of much research. While there is some evidence showing
that place of residence does not affect one's food security status,
(6) the impact that poor geographic access may have on unhealthy eating
patterns should not be underestimated.
Although there is much debate surrounding the exact pathways,
geographic access to food is considered by researchers in the United
States (US) as a risk factor for excess weight and obesity and for some
chronic diseases linked to diet. (7-9) While some argue that proximity
to a source of fresh fruit and vegetables (FV) is not a factor
determining their consumption, (10) studies in the United Kingdom (UK)
and the US have found that coping with poor geographic access to food
may affect the type and quality of food purchased by households with
limited means. (11,12)
A decade of research examining geographic accessibility and food
access was summed up by Beaulac and colleagues in 2009:13 food deserts
are consistently found in American cities, yet Canada, the UK and other
Western countries show no systematic evidence of a correlation between
poverty and poor geographic access to food. On the contrary, several
Canadian studies conclude that lower income areas tend to have better
accessibility than higher income areas. (14-16) Despite this evidence of
overall better geographic access to food for the underprivileged in
Canada, some neighbourhoods where access to food is poor and deprivation
is high have been identified in cities such as Edmonton (AB) and London
(ON). In addition, recent work linking poor food access with certain
urban forms (17) and municipal zoning practices, (18) coupled with a
growing concern over the decentralization of poverty in Canada, (19)
establish the need for more empirical research into disparities in food
access. Accordingly, this paper presents the results of a case study
undertaken in Gatineau, Quebec, to establish whether there are areas
where high deprivation coincides with poor geographic access to food
stores.
The study area: Gatineau, Quebec
Since the 1960s, the city of Gatineau has developed as a
low-density bedroom suburb of Ottawa (ON). It has a population of
approximately 243,000 clustered around five formerly independent
municipalities. In the past decade, Gatineau has expanded in a form
typical of Canadian cities with populations under a million inhabitants:
low-density suburban sprawl. (20) In 1999, the only supermarket in the
central neighbourhood of Hull closed down. Supermarket closure in the
city core is cited as one of the main processes leading to the formation
of food deserts. (15,16) This observation raised the question: are there
problematic areas in Gatineau and if so, can they be described based on
their level of geographic accessibility?
METHODS
The three main components of the study are as follows: 1) the
characterization of the local food environment; 2) the creation of a
local deprivation index; and 3) the measurement of geographic
accessibility to the food environment. These components are brought
together using a k-means cluster analysis and an overlay technique.
Local food environment
As in the study by Bertrand and colleagues, (21) the retail food
stores included in this study have a minimum of seven square metres (75
sq. ft.) of shelf and floor space devoted to the sale of fresh FV. This
excludes convenience and other small stores selling lemons and onions
but not representing a substantial source of fresh FV. A list of food
retailers was obtained from the Ministere de l'Agriculture, des
Pecheries et de l'Alimentation du Quebec (MAPAQ) and crossvalidated
using the Yellow pages. Each store was visited to measure the FV surface
area and a handheld GPS was used to locate the stores on the map.
Forty-five food retailers were retained for this study: 30 supermarkets,
8 grocery stores, 5 non-traditional food retailers, and 2 health food
stores.
Deprivation index
Deprivation indices are common in public health research as they
bring together a number of variables capturing a complex socioeconomic
portrait. The present study combines five variables from the 2006
Canadian census into a composite index of social and material
deprivation (see Table 1). The selection of these variables is inspired
by the work of Pampalon and Raymond, (22) who validated the relationship
between these variables and deprivation. Similar indices have been used
in previous work examining geographic access to food stores in Canada.
(14,16) The dissemination area (DA) was chosen as the unit of
measurement since it is the finest scale at which socio-economic data
are provided by Statistics Canada. The proportions for each variable
were scaled between 0 and 1 and then summed to create the index. The
values for the employment rate and household income were inverted to
insure that they vary in accordance to deprivation; i.e., as income
decreases, deprivation increases. The resulting composite index was
divided into population quartiles; i.e., each quartile contains areas
comprising 25% of the population of the study area (Table 1).
Accessibility measures
Potential geographic accessibility is measured to determine to what
extent individuals residing in a specific DA have access to a given set
of retailers. From the literature, we selected two measures representing
proximity to a source of healthy food: a) the distance (in metres) to
the nearest supermarket and b) the distance to the nearest food store of
any type, including supermarkets. Two other measures representing
variety of food offered within a walkable distance were also selected:
c) the number of retail food stores and d) the total area devoted to the
sale of fresh FV (in m2), both calculated for a 1 km radius from the DA
centroid. To limit aggregation errors, (23) all accessibility measures
were initially computed for each census block (CB) using street network
distances. (24) Each measure was then weighted by the population of the
CB and aggregated to the DA scale (see Apparicio and colleagues (14) for
equations).
[FIGURE 1 OMITTED]
Cluster analysis
All four measures were integrated in the calculation of a k-means
classification, a procedure creating mutually exclusive classes
representing different "types" of accessibility. (25) After
standardizing the variables, the procedure attributes each DA to a class
based on the similarity of its characteristics to the other DAs, while
at the same time the procedure seeks to create classes that are
significantly different from each other. (26)
In addition to Spearman correlations calculated between the
accessibility measures and the deprivation index, a Chi-square test was
undertaken between accessibility classes and deprivation index
quartiles. Finally, high deprivation areas (fourth quartile) were
overlaid on a map with areas classified as having poor and intermediate
accessibility.
RESULTS
After running a series of k-means clustering procedures (4 to 15
classes), the result from the procedure with six accessibility classes
was retained as it had the highest pseudo-F statistic and cubic
clustering criterion. (25) Table 2 presents average values for each
class. For the three classes defined as having poor accessibility, the
average distance to the nearest food store, including supermarkets, is
1.4 km or greater and FV variety within 1 km is limited (median=0 for
both measures). Consumers living in intermediate areas are closer to
other retailers than to supermarkets (632 metres compared to 2 km,
respectively) and have on average only one store within walking distance
and a limited FV availability nearby (16 m2). In the good areas,
consumers are closer to the nearest supermarket (820 metres) and even if
the average FV shelf space is greater (80 m2), they still have access to
only one store within 1 km. Finally, the areas with excellent
accessibility have two or more stores nearby (mean distance close to 600
metres for both measures) and a wide variety of FV (a mean of 2.4 stores
and more than 200 m2 of FV within 1 km).
Gatineau's food gaps
More than 70% of Gatineau's poor currently live within walking
distance of a retail food store selling fresh FV (Table 3). In fact, as
deprivation increases, so does geographic accessibility to a healthy
diet (p<0.0001 for all measures), which is illustrated by negative
correlations between deprivation and the mean distances (-0.248 and
-0.465 for the nearest supermarket and nearest store, respectively), and
positive correlations between deprivation and the variety indicators
(0.421 and 0.311 for number of stores and FV shelf space, respectively).
Conversely, Table 3 shows that almost 30% of the most deprived DAs (Q4)
have poor accessibility (Poor I and Poor II classes combined). These 32
DAs are inhabited by more than 18,000 people or 7.5% of the overall
population of the city. Another 27.5% of Gatineau's deprived DAs -
representing an additional 15,000 people or 6.4% of the overall
population--have intermediate accessibility. Furthermore, the
intermediate class itself is characterized by its association with
deprivation as two thirds of all DAs in this type of environment are
classified as deprived (30 out of 45). To illustrate these results,
Figure 2 locates Gatineau's problematic areas by superimposing
deprived areas with both poor and intermediate accessibility. This map
shows the dispersion of such areas throughout the urbanized sectors.
[FIGURE 2 OMITTED]
DISCUSSION
Long-term exposure to stress and anxiety from living in an area
with high levels of deprivation has been linked to deteriorating health
at the individual level. (27) Add to this the cost and stress of coping
with poor accessibility to a food store and the result is that those
individuals living at or near the poverty line are at a greater risk of
being pushed periodically (or permanently) into unhealthy eating
patterns. (3,4) If, indeed, the geography of food access makes certain
individuals living in food gaps more vulnerable to external factors
(e.g., rise in gasoline prices) and to changes in their personal
circumstances (e.g., change in employment status), then our results
indicate areas where unconventional means may be needed to address food
insecurity--such as the creation of food cooperatives, initiatives to
increase the shelf space devoted to the sale of FV in existing stores,
(28) farmers' markets (29) or urban gardening initiatives.
According to our results, areas with poor accessibility tend to be
dispersed just outside the cores of the former municipalities of
Gatineau; a pattern similar to that in Edmonton, AB. (15) A look at the
morphological characteristics of these areas shows low-density
inner-ring suburban areas intermingled with higher-density residential
housing projects and even a mobile home park. Furthermore, according to
the hypothesis of poverty decentralization in Canadian cities, rising
proportions of low-income households are settling in low-density
inner-ring suburbs. (19) The combination of these two elements may lead
to an increase in the number and scope of Gatineau's food gaps and
may also contribute to an increase in deprivation in existing food gaps.
The areas classified as intermediate food environments are found in
the deprived areas at the centres of the original municipalities. It is
difficult to predict how these areas will evolve over time; on the one
hand, they appear to be at risk of becoming food gaps in the future,
(15,16) yet on the other hand, city projects to revitalize former
"main streets" may improve accessibility levels in these areas
over time. Either way, in the intermediate areas, where the nearest
supermarket is more than 2 km away, any small grocery store closure (or
new store opening, for that matter) or even changes in the type of food
sold in them will modify the foodscape. (30)
Finally, it should be noted that this study only considers
potential geographic accessibility rather than measuring individual
shopping behaviour. Further research needs to be undertaken to answer to
what extent living in an area with poor geographic access actually
affects food-purchasing behaviour and health and well-being. Answers to
these questions, combined with our results, will help planners improve
local food security policies.
Acknowledgements: Mario Corbeil and Philippe Garvie, Direction de
Sante publique de l'Outaouais.
Conflict of Interest: None to declare.
Received: February 1, 2012 Accepted: September 3, 2012
REFERENCES
(1.) Raphael D. The social determinants of health: An overview of
key issues and themes. In: Raphael D (Ed.), The Social Determinants of
Health: Canadian Perspectives, 2nd ed. Toronto, ON: Canadian
Scholars' Press Inc., 2009.
(2.) White HL, Matheson FI, Moineddin R, Dunn JR, Glazier RH.
Neighbourhood deprivation and regional inequalities in self-reported
health among Canadians: Are we equally at risk? Health & Place
2011;17(1):361-69.
(3.) McIntyre L. Food security: More than a determinant of health.
Policy Options 2003; March: 46-51.
(4.) Office of Nutrition Policy and Promotion. Canadian Community
Health Survey, Cycle 2.2, Nutrition (2004): Income-Related Household
Food Security in Canada. Ottawa, ON: Health Canada, 2007.
(5.) Shaw HJ. Food deserts: Towards the development of a
classification. Geografiska Annaler Series B-Human Geography
2006;88B(2):231-47.
(6.) Kirkpatrick SI, Tarasuk V. Assessing the relevance of
neighbourhood characteristics to the household food security of
low-income Toronto families. Public Health Nutr 2010;13(07):1139-48.
(7.) 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(4):333-39.
(8.) Stafford M, Cummins S, Ellaway A, Sacker A, Wiggins RD,
Macintyre S. Pathways to obesity: Identifying local, modifiable
determinants of physical activity and diet. Soc Sci Med
2007;65(9):1882-97.
(9.) Ewing RT, Schmid T, Killingsworth R, Zlot A, Raudenbush S.
Relationship between urban sprawl and physical activity, obesity, and
morbidity. In: Marzluff JM, Bradley G, Shulenberger E, et al. (Eds.),
Urban Ecology. Seattle, WA: Springer US, 2008.
(10.) Pearson T, Russell J, Campbell MJ, Barker ME. Do 'food
deserts' influence fruit and vegetable consumption? -A
cross-sectional study. Appetite 2005;45(2):195-97.
(11.) Moore LV, Diez Roux AV, Nettleton JA, Jacobs DR. Associations
of the local food environment with diet quality--A comparison of
assessments based on Surveys and Geographic Information Systems: The
Multi-Ethnic Study of Atherosclerosis. Am J Epidemiol
2008;167(8):917-24.
(12.) Whelan A, Wrigley N, Warm D, Cannings E. Life in a 'food
desert'. Urban Studies 2002;39(11):2083-100.
(13.) Beaulac J, Kristjansson E, Cummins S. A systematic review of
food deserts, 1966-2007. Prev Chron Dis 2009;6(3):1-10.
(14.) 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.
(15.) Smoyer-Tomic KE, Spence JC, Amrhein C. Food deserts in the
prairies? Supermarket accessibility and neighborhood need in Edmonton,
Canada. Professional Geographer 2006;58(3):307-26.
(16.) Larsen K, Gilliland J. Mapping the evolution of 'food
deserts' in a Canadian city: Supermarket accessibility in London,
Ontario, 1961-2005. Int J Health Geogr 2008;7:16.
(17.) Daniel M, Kestens Y, Paquet C. Demographic and urban form
correlates of healthful and unhealthful food availability in Montreal,
Canada. Can J Public Health 2009;100(3):189-93.
(18.) Black JL, Carpiano RM, Fleming S, Lauster N. Exploring the
distribution of food stores in British Columbia: Associations with
neighbourhood sociodemographic factors and urban form. Health &
Place 2011;17(4):961-70.
(19.) Ades J, Apparicio P, Seguin A-M. Are new patterns of
low-income distribution emerging in Canadian metropolitan areas? The
Canadian Geographer/Le Geographe canadien 2012;56(3):339-61.
(20.) Filion P, Bunting T, Pavlic D, Langlois P. Intensification
and sprawl: Residential density trajectories in Canada's largest
metropolitan regions. Urban Geogr 2010;31(4):541-69.
(21.) Bertrand L, Therien F, Cloutier M-S. Measuring and mapping
disparities in access to fresh fruits and vegetables in Montreal. Can J
Public Health 2008;99(1):6-11.
(22.) Pampalon R, Raymond G. A deprivation index for health and
welfare planning in Quebec. Chron Dis Can 2000;21(3):104-13.
(23.) Hewko J, Smoyer-Tomic KE, Hodgson MJ. Measuring neighbourhood
spatial accessibility to urban amenities: Does aggregation error matter?
Environ Plan ning2002;34(7):1185-206.
(24.) Apparicio P, Abdelmajid M, Riva M, Shearmur R. Comparing
alternative approaches to measuring the geographical accessibility of
urban health services distance types and aggregation-error issues. Int J
Health Geogr 2008;7:7.
(25.) Everitt BS, Landau S, Leese M. Cluster Analysis. London, UK:
Edward Arnold, 1997.
(26.) SAS Institute Inc. SAS version 9.2. Cary, NC, USA.
(27.) Giordano GN, Lindstrom M. The impact of changes in different
aspects of social capital and material conditions on self-rated health
over time: A longitudinal cohort study. Soc Sci Med 2010;70(5):700-10.
(28.) Morland KB. An evaluation of a neighborhood-level
intervention to a local food environment. Am J Prev Med
2010;39(6):31-38.
(29.) Larsen K, Gilliland J. A farmers' market in a food
desert: Evaluating impacts on the price and availability of healthy
food. Health & Place 2009;15(4):115862.
(30.) Institute of Medicine and National Research Council. The
Public Health Effects of Food Deserts: Workshop Summary. Washington, DC:
National Academies Press, 2009.
Adrian C. Gould, MSc, Philippe Apparicio, PhD, Marie-Soleil
Cloutier, PhD
Author Affiliations
Centre Urbanisation Culture Societe, Institut national de la
recherche scientifique, Montreal, QC
Correspondence: Marie-Soleil Cloutier, Centre Urbanisation Culture
Societe, Institut national de la recherche scientifique, 385 rue
Sherbrooke Est, Montreal, QC H2X 1E3, Tel: 514-499-4096, Fax:
514-499-4065, E-mail: marie-soleil.cloutier@ucs.inrs.ca
Table 1. Mean Values of Deprivation-related Variables and Deprivation
Index According to Population Quartile
Population Quartile
(Mean Value)
Q1 Q2
(Low Deprivation)
Proportion of separated, 12.5 16.8
divorced or widowed
persons (%)
Proportion of single-parent 10.7 15.9
families (%)
Proportion of individuals (%) 6.0 11.6
aged 24 to 65 without a
high school diploma
Employment rate (%) 76.1 71.5
Median household income $80,388 $58,383
before taxes ($)
Deprivation index 1.05 1.59
Population Quartile
(Mean Value)
Q3 Q4 All
Dissemination
(High Areas
Deprivation)
Proportion of separated, 19.8 24.8 19.1
divorced or widowed
persons (%)
Proportion of single-parent 21.6 30.2 20.6
families (%)
Proportion of individuals 18.2 29.9 17.6
aged 24 to 65 without a
high school diploma
Employment rate (%) 64.2 54.9 65.6
Median household income $44,917 $32,534 $51,283
before taxes ($)
Deprivation index 2.10 2.81 1.98
Table 2. Mean Values for Each Accessibility Measure
According to Accessibility Class
Accessibility Class Nearest Nearest Store
(Number of DA) Supermarket * ([dagger])
Very poor (n=13) 5371 5329
Poor II (n=63) 2802 2637
Poor I (n=140) 1443 1358
Intermediate (n=46) 2038 632
Good (n=94) 820 777
Excellent (n=36) 627 597
All classes (n=392) 1637 1637
FV Shelf
Accessibility Class Store Variety Space
(Number of DA) ([double dagger]) ([section])
Very poor (n=13) 0.00 0.00
Poor II (n=63) 0.00 0.00
Poor I (n=140) 0.00 0.00
Intermediate (n=46) 1.00 16.00
Good (n=94) 1.00 80.50
Excellent (n=36) 2.36 224.00
All classes (n=392) 0.33 10.50
* The mean distance to the nearest supermarket (metres).
([dagger]) The mean distance to the nearest food store of any
type (metres).
([double dagger]) The median number of stores within a walkable
distance of 1000 m.
([section]) The median FV shelf space in stores within a
distance of 1000 m ([m.sup.2]).
Table 3. The Proportion of Dissemination Areas in Each Deprivation
Quartile Broken Down by Accessibility Class
(% of the Number of DAs in Each Class)
Accessibility Class
Deprivation Index
Population Quartile Very Poor Poor II Poor I
Q4 (high deprivation) 0.0 4.6 24.8
Q3 1.8 10.6 41.6
Q2 3.2 24.5 39.4
Q1 (low deprivation) 11.1 30.6 38.9
Accessibility Class
Deprivation Index
Population Quartile Intermediate Good Excellent
Q4 (high deprivation) 27.5 29.4 13.8
Q3 11.5 26.5 8.0
Q2 2.1 23.4 7.4
Q1 (low deprivation) 0.0 12.5 6.9
Chi-square = 95.812 (p<0.0001); Pearson
correlation = 0.359 (p<0.0001).
