Vol.1, No.3, 182-189 (2011)
doi:10.4236/ojpm.2011.13024
C opyright © 2011 SciRes. Openly accessible at http://www.scirp.org/journal/OJPM/
Open Journal of Preventive Medicine
Neighbourhood differences in objectively measured
physical activity, sedentary time and body mass index
Stephanie A. Prince1,2*, Mark S. Tremblay2,3,4, Denis Prud’homme3, Rachel Colley2,4,
Michael Sawada5, Elizabeth Kristjansson6
1Population Health PhD Program, University of Ottawa, Ottawa, Canada; *Corresponding Author:
2Healthy Active Living and Obesity Research Group, Children’s Hospital of Eastern Ontario, Ottawa, Canada;
3Faculty of Health Sciences, University of Ottawa, Ottawa, Canada;
4Faculty of Medicine, University of Ottawa, Ottawa, Canada;
5Laboratory for Applied Geomatics and GIS Science (LAGGISS), Department of Geography, University of Ottawa, Ottawa, Canada;
6School of Psychology, University of Ottawa, Ottawa, Canada.
Received 9 September 2011, revised 13 October 2011; accepted 23 October 2011.
ABSTRACT
Background: There is limited Canadian re-
search examining whether directly measured
physical activity (PA) and body mass index
(BMI) differ between neighbourhoods with dif-
ferent objectively measured socioeconomic
(SES) and recreation (REC) environments. Pur-
pose: To determine whether mean adult PA
levels, sedentary time and BMIs were different
across four neighbourhoods with contrasting
SES and REC environments in Ottawa, Can-
ada. Methods: This study employed a cross-
sectional design to collect pilot data of objec-
tively measured height, weight and PA (using
accelerometry) and self-reported covariates in
113 adults (≥18 years). Four contrasting neigh-
bourhoods (high REC/high SES, high REC/low
SES, low REC/high SES, and low REC/low SES)
were selected based on data collected as part
of the Ottawa Neighbourhood Study. Analysis
of covariance and logistic regression were
used to perform neighbourhood comparisons
for PA, sedentary time and BMI, adjusting for
age, sex and household income and possible
interactions. Post-hoc comparisons using Tu-
key’s test were performed. Results: Signifi-
cant neighbourhood-group effects were ob-
served for light intensity PA and sedentary
time. Post-hoc tests identified that the low
REC/high SES neighbourhood had significant-
ly more minutes of light PA than the low
REC/low SES (Mdiff = 56.05 minutes·day, Tukey
p = 0.01). Unadjusted BMI differed between the
four neighbourhoods, but the differences were
not significant after controlling for age, sex
and household income. Conclusions: This
study demonstrates that light PA and seden-
tary time differ between neighbourhoods of
varying REC and SES environments after con-
trolling for differences in age, sex and house-
hold income. Findings also suggest that other
area-level factors may explain these neigh-
bourhood differences.
Keywords: Physical Activity; Sedentary Time;
Obesity; Neighbourhood; Environment
1. INTRODUCTION
Despite the benefits of daily physical activity (PA)
and a healthy body weight for the prevention of several
chronic diseases and premature mortality, most adult
Canadians continue to fall short of PA recommendations
and are overweight or obese [1-3]. Evidence suggests a
link may exist between the recreation and social envi-
ronments and an individual’s likelihood of being physi-
cally active or overweight/obese [4-6]. Historically, the
majority of research in this area has focused on per-
ceived access to environments (e.g. what an individual
feels they have access to or what they perceive their en-
vironment to be like) rather than objective measures of
their built and socioeconomic environment (e.g. the
number of facilities in their neighbourhood or average
income levels) [7]. Results from objectively measured
studies are mixed, but generally report positive associa-
tions between greater access to recreation environments
and PA while lower access tends to be associated with
greater odds of overweight and obesity [7-10]. Impor-
tantly, fewer studies have linked objectively measured
environments to directly measured PA, sedentary time
and body mass index (BMI) and even fewer have done

S. A. Prince et al. / Open Journal of Preventive Medicine 1 (2011) 182-189
Copyright © 2011 SciRes. Openly accessible at http://www.scirp.org/journal/OJPM/
183
so in a Canadian context [7,10]. While some research
has examined the independent relationship of either the
recreation or social environments with PA and BMI, less
is known about the possible synergistic effects of these
two determinants. Research looking at environmental
factors associated with light PA and sedentary time is
emerging, but remains limited. In Canada, availability of
recreation resources may be lower among neighbour-
hoods of lower socioeconomic status (SES) and, as such,
an area’s SES may affect the relationship of the recrea-
tion environment on PA [11-14].
In light of the possible effect of neighbourhood SES
on the relationship between the recreation environment
and PA and BMI, and current gaps in the literature, the
main aim of this study was to compare objectively mea-
sured mean PA levels, sedentary behaviour and BMIs
across four neighbourhoods with contrasting SES and re-
creation environments in adults living in Ottawa, Canada.
Our hypothesis was that the neighbourhood with the
fewest recreation resources and lowest neighbourhood
SES would have the lowest levels of PA, greatest a-
mounts of sedentary time and largest BMIs.
2. METHODS
This study was carried out in Ottawa, a large Cana-
dian city with a regional population of approximately 1.2
million residents. Mean minutes of directly measured
light, moderate and vigorous PA, sedentary time, step
counts, and BMI were compared between participants
living in four pre-selected, contrasting neighbourhoods
characterized by their objectively measured SES and
recreation (REC) environments. The study received
ethical approval from the University of Ottawa’s Health
Science Research Ethics Board and the Children’s Hos-
pital of Eastern Ontario (CHEO) Research Ethics Board.
2.1. Participants
The study population includes 113 adults (≥18 years)
with complete PA and BMI data. Participants were re-
cruited through community newsletters, bulletin board
posters, mailbox flyers, mail postcards, and word of
mouth. Respondents were required to speak and read
English or French, agree to wear an accelerometer for a
minimum of five weekdays and two weekend days, have
their heights and weights measured, and live within the
boundaries of one of the study neighbourhoods as de-
termined by their residential address. One adult per
household was eligible to participate. Participants were
provided with a $20 honorarium and a PA profile fol-
lowing participation in the study. Study measurement
sessions took place in public facilities where written
informed consent was obtained, height and weight mea-
surements were taken and instructions for accelerometer
use were explained. Participants were provided with an
Actical® accelerometer (Phillips-Respironics, Oregon,
USA), an instruction booklet/questionnaire, and a pre-
paid addressed envelope to return the study materials
and accelerometer.
2.2. Study Neighbourhoods
Participants were recruited from four neighbourhoods
within the City of Ottawa. The neighbourhood bounda-
ries were defined by the Ottawa Neighbourhood Study
(ONS; www. neighbo urhoodst udy.ca), a large study
examining associations between neighbourhood charac-
teristics and health outcomes. Figure 1 provides a map
of the ONS neighbourhoods with the four chosen study
neighbourhoods highlighted. The neighbourhoods were
chosen to represent four contrasting areas based on REC
and SES environments. The neighbourhoods are as fol-
lows: 1) high REC/high SES; 2) high REC/low SES; 3)
low REC/high SES; and 4) low REC/low SES.
2.3. Ottawa Neighbourhood Study (ONS)
In the ONS, neighbourhoods were defined based on
natural boundaries, similarity in SES and demographics,
Ottawa Multiple Listing Services maps, and participatory
mapping feedback from community members and experts
[15]. Objectively measured environmental data were col-
lected from 2006 to 2008 using the following data and
methods: 1) 2006 Canadian census household data; 2) GIS
data from DMTI Spatial Inc., the City of Ottawa, and the
National Capital Commission; 3) telephone contact with
businesses; 4) web-based research; 5) team knowledge of
local resources; and 6) field research and validation. A
further in-depth description of methods related to the ONS
and its variables is available elsewhere [15].
2.4. Recreation (REC) Environment
The REC environment was based on a REC Index that
was created using principal components analysis; this
served as a measure of the density of facilities for rec-
reation in a neighbourhood [15]. The REC index in-
cludes meters of bike and walking paths per person, me-
ters-squared of park space per person, and recreation
facilities per thousand people per neighbourhood. The
REC index was t-scored to represent a mean of 50 with a
standard deviation of 10 for comparability across neigh-
bourhoods. Recreational facilities were defined using the
North American Industry Classification System—Cana-
da (NAICS) Code 71 and were only included if they
were free or had a minimal cost [16]. Green space man-
aged by the City or National Capital Commission was
ncluded in the park variable. i

S. A. Prince et al. / Open Journal of Preventive Medicine 1 (2011) 182-189
Copyright © 2011 SciRes. http://www.scirp.org/journal/OJPM/
184
Figure 1. Ottawa neighbourhood study map and the four study neighbourhoods.
2.5. Socio-Economic (SES) Environment
The SES environment was assessed using a neigh-
bourhood SES index developed using principal compo-
nents analysis; it included percent of households below
the low-income cut-off (LICO), average household in-
come, percent of unemployed residents, percent of resi-
dents with less than a high school education, and percent
of single-parent families [15,17]. The SES index was
t-scored to represent a mean of 50 with a standard devia-
tion of 10.
Openly accessible at
2.6. Individual-Level Outcomes
2.6.1. Physical Activity (PA) and Sedentary Time
PA was directly measured using Actical® acceler-
ometers (Phillips-Respironics, Oregon, USA). The ac-
celerometers are small, omni-directional, water resistant
movement sensors able to capture all intensities of
movement, including sedentary pursuits. The Actical®
has been validated to measure PA and step counts in
adults [18,19]. Participants were asked to wear the ac-
celerometers on their right hip using an elasticized belt
during their waking hours for a total of five week days
and two weekend days. At midnight after the study
meeting, data recording commenced. An epoch length of
60-seconds was used to capture movement in a count
value per minute (cpm). Signals were also translated into
steps per minute.
Valid accelerometer results were selected based on
guidelines adopted for the data analyses of the Canadian
Health Measures Survey (CHMS) [2,20]. A valid day
was defined as ≥10 hours of wear time and participants
were required to have ≥4 valid days to be retained for
the analyses. Data reduction and analysis was harmo-
nized with the CHMS accelerometry data procedures.
Further detail is available elsewhere [2,20].
The accelerometer data were downloaded using the
Actical® software program and analysed in SAS. Data
are presented as average daily minutes spent in moderate
PA, vigorous PA, moderate-to-vigorous PA (MVPA),
light activity, and sedentary time, and mean daily steps.
PA intensity cut-points used in the identification of ac-
tivity levels are provided in Tabl e 1. PA was also ana-
lysed to assess those meeting current Canadian PA
Guidelines and those who obtained an average of 10,000
steps per day versus those that did not [21,22]. If a par-
ticipant had between four-to-six valid days, their weekly
sum was calculated by multiplying the mean daily
MVPA by seven.
2.6.2. Body Mass Index (BMI)
BMI was calculated using directly measured weight
(kg) divided by height squared (m2). Standing height was
measured using a Seca 214 portable stadiometer (SECA,
Hanover MD, USA) and recorded to the nearest centi-
meter and converted to meters. Weight was measured
using a Life Source ProFit scale (A&D Medical,
Milpitas, CA, USA) and recorded to the nearest 0.1 kg.
BMI guidelines for adults were used to group individuals
into the following categories: underweight (<18.5

S. A. Prince et al. / Open Journal of Preventive Medicine 1 (2011) 182-189
Copyright © 2011 SciRes. Openly accessible at http://www.scirp.org/journal/OJPM/
185
Table 1. Physical activity intensity cut-points for the actical as used by the Canadian Health Measures Survey [2,23,24].
Intensity Metabolic Equivalent (METS) Example Accelerometer count
range (CPM)
Sedentary 1 to less than 2 Car travel, sitting, reclining, standing Less than 100*
Light 2 to less than 3 Walking less than 3.2 km/h, light household cleaning, cooking 100 to less than 1535
Moderate 3 to less than 6 Walking less than 3.2 km/h, cleaning (vacuuming, washing car),
bicycling for pleasure 1535 to less than 3962
Vigorous 6 or more Jogging, competitive team sport participation 3962 or more
CPM—counts per minute *including wear-time zeros.
kg·m2), normal weight (18.5 - 24.9 kg·m2), overweight
(25.0 - 29.9 kg·m2), and obese (30 kg·m2) for descrip-
tive purposes only [25]. BMI was analysed as a con-
tinuous outcome.
2.6.3. Individual-Level Covariates
Age and sex were forced into the models based on
their known associations with PA and BMI. Household
income was included based on its significant bivariate
association with PA. Information on work, education,
marital and smoking status and number of individuals
per household were also collected, but were all signifi-
cantly correlated with household income. To avoid mul-
ticollinearity and obtain the most parsimonious model
given the sample size, they were not included. Season of
data collection was also available; however, it was not
significantly associated with either PA or BMI. All co-
variates were self-reported. Missing information for age
or income was imputed based on regression parameter
estimates from models using sex, neighbourhood and
BMI as predictors. In total, three participants were
missing information on age and an additional seven were
missing information for income.
Age was included as a continuous measure. Sex was
coded as male or female. Household income was con-
sidered as a two-level categorical variable that accounts
for the number of people in the household and total
household income from all sources in the 12 months
before the interview: “lower” (<$60,000 for 1 - 2 people;
<$80,000 for ≥3 people) compared to “upper” (≥$60,000
for 1 - 2 people; ≥$80,000 for ≥3 people).
2.7. Statistical Analysis
All analyses were conducted using SAS version 9.2
(SAS Institute Inc., Cary, NC, USA). Unadjusted de-
scriptive statistics were performed to derive means ±
standard deviations and frequencies of all demographic
variables for each neighbourhood. Unadjusted between
neighbourhood comparisons were analysed using one-
way analysis of variance (ANOVA) for continuous vari-
ables and chi-square tests for dichotomous variables.
Adjusted between neighbourhood comparisons were
performed using analysis of covariance (ANCOVA) for
continuous outcomes (using Proc GLM and LSMeans
for unequal designs) and logistic regression for binary
outcomes (using Proc Logistic and Proc GLM). AN-
COVA models allowed for the adjustment of differences
in the covariates between neighbourhoods and between
subjects in each neighbourhood. Interactions were tested
between all covariates and neighbourhoods with no sig-
nificant interactions observed. Post-hoc Tukey tests were
used to perform pairwise comparisons for PA and BMI
between the individual neighbourhoods.
3. RESULTS
3.1. Sample Characteristics
A total of 126 adults (≥18 years) were recruited from
four neighbourhoods. From this sample, 113 had com-
plete BMI and valid PA data and were used in the analy-
ses. Table 2 provides descriptive characteristics of the
participants by neighbourhood.
The participants were reasonably well distributed a-
cross the neighbourhoods. Compared to the general Ca-
nadian population as assessed by the CHMS, the study
participants were leaner than a representative Canadian
sample (53% versus 38% healthy weight), had similar
proportions meeting step-per-day guidelines (35% versus
34%) and had a greater proportion of individuals meeting
MVPA guidelines (31% versus 15%) [2,3].
3.2. Neighbourhood Differences in Physical
Activity
Tables 3 provides comparisons of mean PA levels and
proportions meeting the current Canadian PA guidelines
[21]. Unadjusted comparisons showed that only mean
daily minutes spent in light PA (p = 0.01) differed be-
tween participants living in the four neighbourhoods;
this difference was stronger after adjustment for co-
variates (p = 0.002). Following adjustment, a signifi-
cant group effect was observed for minutes of seden-
tary behaviour and light PA. The low REC/high SES
neighbourhood had the most minutes spent engaged in
light PA with post-hoc tests identifying average min-
utes were significantly higher than the low REC/low
SES neighbourhood (Mdiff = 56.05 minutes·day, Tukey
p = 0.01).
Following adjustment, mean differences in average

S. A. Prince et al. / Open Journal of Preventive Medicine 1 (2011) 182-189
Copyright © 2011 SciRes. Openly accessible at http://www.scirp.org/journal/OJPM/
186
Table 2. Participant characteristics by neighbourhood (n = 113).
Characteristics High REC/ High
SES (n=29)
High REC/
Low SES (n=29)
Low REC/
High SES (n=23)
Low REC/
Low SES (n=32)
Male sex, n (%) 8 (28%) 6 (21%) 4 (17%) 7 (22%)
Age, years 51.3 ± 13.7 48.1 ± 14.1 51.0 ± 14.1 45.8 ± 15.0
BMI category, n (%)
Underweight 0 1 (3%) 0 0
Healthy weight 17 (59%) 19 (66%) 7 (31%) 17 (53%)
Overweight 9 (31%) 7 (24%) 12 (52%) 8 (25%)
Obese 3 (10%) 2 (7%) 4 (17%) 7 (22%)
Household income, n (%)*
Low 2 (7%) 7 (24%) 7 (30%) 18 (56%)
High 27 (93%) 22 (93%) 16 (70%) 14 (44%)
Data are presented as mean ± standard deviations unless otherwise stated. REC – recreation index score, SES – socioeconomic status index score. *Mean
differences between neighbourhoods based on chi-square or ANOVAs p < 0.001.
Tab le 3. Unadjusted and adjusted* average daily and weekly minutes of activity at various levels of intensity, average daily step
counts, percent meeting physical activity guidelines, and body mass index by neighbourhood (total n = 113).
Unadjusted Adjusted*
High REC/
High SES
(n=29)
High REC/
Low SES
(n=29)
Low REC/
High SES
(n=23)
Low REC/
Low SES
(n=32)
p-value
High REC/
High SES
(n=29)
High REC/
Low SES
(n=29)
Low REC/
High SES
(n=23)
Low REC/
Low SES
(n=32)
p-value
Activity intensity
Daily minutes of sedentary 581 ± 13 602 ± 13 570 ± 15596 ± 130.39 591 ± 16611 ± 15 576 ± 16 599 ± 13 0.05
Daily minutes of light 98 ± 12 87 ± 12 114 ± 1457 ± 12 0.01 88 ± 14 76 ± 13 105 ± 15 49 ± 12 0.002
Daily minutes of moderate 15 ± 4 13 ± 4 14 ± 4 20 ± 4 0.52 13 ± 4 12 ± 4 13 ± 5 20 ± 4 0.82
Daily minutes of vigorous 4 ± 1 2 ± 1 1 ± 2 3 ± 1 0.67 3 ± 2 2 ± 2 1 ± 2 3 ± 1 0.24
Daily minutes of MVPA 19 ± 4 16 ± 4 17 ± 5 23 ± 4 0.59 17 ± 5 14 ± 5 15 ± 5 23 ± 4 0.82
Weekly minutes of MVPA 134 ± 29 112 ± 29 115 ± 32161 ± 270.59 118 ± 35100 ± 33 105 ± 36 159 ± 30 0.82
Daily step counts 9446 ± 682 8518 ± 682 8596 ± 7668413 ± 6490.69 9208 ± 8158227 ± 768 8436 ± 841 8217 ± 6960.57
Met weekly MVPA
guidelines, n (%) 10 (34%) 9 (31%) 6 (26%) 10 (31%)0.94 8 (31%)8 (27%) 6 (22%) 12 (29%)0.97
Met daily step guidelines, n
(%) 12 (41%) 9 (31%) 8 (35%) 9 (28%) 0.75 10 (38%)12 (27%)6 (32%) 12 (25%)0.34
Body mass index, kg·m2 24.7 ± 0.8 23.9 ± 0.8 27.0 ± 0.926.3 ± 0.80.04524.8 ± 1.024.2 ± 0.9 27.3 ± 1.0 26.7 ± 0.80.13
All activity is reported in average minutes except for step counts. All results shown as mean ± standard error. P-value for ANOVA/ANCOVA or chi-square/
logistic regression models. MVPA—moderate-to-vigorous physical activity. *Adjusted for age, sex, household income.
steps per day remained non-significant between neigh-
bourhoods. Furthermore, no significant differences in the
proportion of participants meeting the current PA guide-
lines either by weekly minutes of MVPA or daily step
counts were observed. The data were also analysed using
percentage of wear time spent in the various levels of PA
in order to adjust for the fact that individuals with longer
wear times may have subsequently higher minutes of
time spent in light and sedentary pursuits. Both the un-
adjusted and adjusted results were virtually identical to
those found for the mean minutes spent at the various PA
intensities.
3.3. Neighbourhood Differences in Body
Mass Index (BMI)
Table 3 provides comparisons of mean BMIs across
the four neighbourhoods. Unadjusted comparisons iden-
tified that BMI differed between participants living in
the four neighbourhoods (p = 0.045); this difference lost
significance following adjustment (p = 0.13). Partici-
pants from the two low REC neighbourhoods had the
highest average BMIs. Post-hoc tests identified that the
low REC/high SES had a higher average BMI than the
high REC/low SES neighbourhood, but this difference
only approached significance (Mdiff = 3.09 kg·m2, Tukey
p = 0.06).
4. DISCUSSION
This study aimed to investigate whether PA levels,
sedentary time and BMI differed between adults living
in four Ottawa neighbourhoods with contrasting SES
and REC environments. To the best of our knowledge,

S. A. Prince et al. / Open Journal of Preventive Medicine 1 (2011) 182-189
Copyright © 2011 SciRes. Openly accessible at http://www.scirp.org/journal/OJPM/
187
this is the first study to use objective measures of the
neighbourhood environment and directly measured PA,
sedentary time and BMI in a Canadian population. Fur-
thermore, it is one of the first to look at objective meas-
ures of the REC environment, rather than previously
examined land-use mix and walkability characteristics,
along with SES on directly measured PA levels, seden-
tary time and BMI. The results support the notion that
PA, sedentary time and BMI may differ by factors in the
built and social environments and that these differences
can be independent of individual-level determinants. The
findings do not support our original hypothesis that indi-
viduals living in a neighbourhood with a combination of
low SES and low availability of recreation facilities
would have the lowest amounts of daily PA, greatest
amount of sedentary time and highest BMIs.
Interestingly, we found that minutes spent in light PA
were significantly greater in the low REC/high SES
neighbourhood compared to the low REC/low SES
neighbourhood. Research on impacts of the environment
on light PA is limited and as mentioned, especially using
objective measurements. Research on a small sample of
Belgian adults (n = 120) demonstrated that adults living
in a ‘high walkable’ neighbourhood took more steps per
day and walked more for transport than those living in a
‘low walkable’ neighbourhood [26]. Research has also
shown that higher mixed-land use is associated with
greater levels of PA, possibly due to increased use of
active transport [8]. While walkability and mix-land use
factors were not used to select our study neighbourhoods,
they do offer a reasonable explanation for why light PA
was higher in the low REC/high SES neighbourhood.
Area zoning measures were calculated for all four
neighbourhoods as a proxy for land use variety. The low
REC/low SES neighbourhood had the greatest number
of different zonings at 114 compared to the high REC/
high SES neighbourhood with 34 and the high REC/low
SES neighbourhood with 30. The low REC/high SES
neighbourhood had 113 different zonings; however, as
can be seen in Figure 1, this neighbourhood is one of the
largest in Ottawa and naturally would have a greater
distribution simply due to its geographical size. The
density of the zonings would therefore be lower than in
the low REC/low SES neighbourhood.
While walkability was not assessed in the current
study, the low REC/high SES neighbourhood is rural
compared to the urban and more walkable low REC/low
SES neighbourhood. Forms of active transportation are
often included as MVPA, therefore, the light PA may be
capturing more incidental movements associated with
lower intensity activities of daily living such as house
and yard work [23]. It is also likely that the low
REC/high SES neighbourhood has larger homes and
properties, possibly resulting in greater time spent in
light household chores. Research has also shown that
possible urban-suburban-rural differences exist for PA
with rural neighbourhoods more likely to be inactive
[27-29]. It is possible that a lower land-mix use and de-
creased walkablity may provide an explanation for why
daily light PA was higher in the low REC/high SES
neighbourhood compared to the low REC/low SES
neighbourhood. Furthermore, research on neighbour-
hood SES has also shown that adults living in areas de-
scribed as lower SES are more likely to use active
transport and less likely to use motorized transport [30].
Research on the impact of neighbourhood environ-
ments on sedentary behaviour is very limited. Recent
work by Van Dyck and colleagues looked at the effect of
neighbourhood walkability on accelerometer measured
sedentary time in adult men and women from Belgium
[31]. Their study identified that contrary to general hy-
potheses, living in a highly walkable neighbourhood was
associated with greater amounts of time spent in seden-
tary pursuits [31]. Interestingly, neighbourhood SES was
not significantly associated with daily inactivity and
neighbourhood SES did not modify the relationship be-
tween walkability and sedentary time [31]. These find-
ings are similar to those seen in the current study
whereby our more urban and walkable neighbourhoods
had higher amounts of sedentary time compared to our
rural and less walkable (low REC/high SES) neighbour-
hood.
In another investigation the current authors examined
the multilevel associations of the recreation and social
environments on self-reported PA and BMI across all of
the neighbourhoods in the City of Ottawa using data
from the Canadian Community Health Survey (Prince
SA, Kristjansson EA, Russell K et al. Relationships
between neighbourhoods, physical activity and obesity:
A multilevel analysis of a large Canadian City.
Unpublished). Findings from that study indicated that
leisure time PA and overweight/obesity were not signifi-
cantly associated with any recreation resources or social
environment variables. In light of these findings and that
our neighbourhoods were only selected based on REC
and SES, it is likely that other environmental character-
istics are influencing PA levels. The relative influence of
where people live and work on their PA and BMI may
differ by population density characteristics and/or other
factors and requires further study. Finally, it is possible
that more robust, direct measures may produce funda-
mentally different results than self-reported measures,
and indeed that convention thinking of the influence of
built and social environments needs to be challenged and
perhaps there are other factors yet to be identi-
fied/measured that are impacting upon the environment’s

S. A. Prince et al. / Open Journal of Preventive Medicine 1 (2011) 182-189
Copyright © 2011 SciRes. Openly accessible at http://www.scirp.org/journal/OJPM/
188
relationship with PA and BMI.
The present study has limitations that should be rec-
ognized. The study’s sample size may have been too
small to see possible differences and clear trends. It must
be noted that recruitment for this study was onerous
likely due to the nature of the direct measurement of PA
and BMI and the constraints placed on location and
times of participant meetings by the Research Ethics
Boards. This likely resulted in a biased sample of highly
motivated individuals, a common bias in research of this
nature. The population was also highly educated (23%
had a graduate-level degree e.g. masters, doctorate) and
affluent (70% had a high household income) and the
results are therefore not generalizable. The study should
therefore be treated as pilot-level information to chal-
lenge other researchers to employ more rigorous meas-
urement methodologies.
It is also important to realize that the accelerometers
provide objective measures that remove self-report bi-
ases, but they are unable to capture all forms of PA ac-
curately (e.g. weight lifting, rowing, cycling). Partici-
pants were also asked to remove the monitors during
water activities and as a result swimming was not cap-
tured in the PA data.
5. CONCLUSIONS
This is the first known Canadian study to examine
whether objectively measured neighbourhood recreation
and SES environments are associated with levels of di-
rectly measured PA and BMI. Results of this study iden-
tify that minutes of light intensity PA and sedentary be-
haviour differ between neighbourhoods of varying rec-
reation and SES environments after controlling for dif-
ferences in age, sex and income. Findings also suggest
that other area-level factors such as mixed-land use and
degree of urbanism may explain these neighbourhood
differences. Future research into the impact of the
neighbourhood on PA and BMI should concomitantly
assess multiple aspects of the environment including, but
not limited to recreation, amenities, walkability, mixed-
land use, population density, SES, social support, social
cohesion, crime, and perceptions and preferences of the
neighbourhood.
6. ACKNOWLEDGEMENTS
The authors are grateful to Travis Saunders for his efforts in partici-
pant recruitment, data collection and verification and to Dr. Jean-Mi-
chelle Billette and Ms. Megan Carter for their statistical and methodo-
logical advice. Support for the ONS was provided by the Canadian
Institutes of Health Research (funding number 99345), the Champlain
Local Health Integration Network, the Ottawa Coalition of Community
Health, Resource Centres and United Way Ottawa. Funding for the
collection of individual data was provided by a Faculty of Health Sci-
ences and CHEO Partnership Research Grant. S. Prince received fund-
ing in support of her doctoral work from the Social Sciences and Hu-
manities Research Council of Canada-Doctoral Award, Ontario Minis-
tries of Ontario Graduate Scholarships and a University of Ottawa
Excellence Scholarship and Doctoral Research Award.
REFERENCES
[1] Katzmarzyk, P.T. and Janssen, I. (2004) The economic
costs associated with physical inactivity and obesity in
Canada: An update. Canadian Journal of Applied Phy-
siology, 29, 90-115. doi:10.1139/h04-008
[2] Colley, R.C., Garriguet, D., Janssen, I., Craig, C.L.,
Clarke, J. and Tremblay, M.S. (2011) Physical activity of
Canadian adults: Accelerometer results from the 2007 to
2009 Canadian Health Measures Survey. Health Reports,
22, 7-14.
[3] Canadian Health Measures Survey: Cycle 1 data tables
(2011) Table 35: Distribution of the household po-
pulation aged 18 to 79, by body mass index norms based
on measured inputs, by age and sex, Canada, 2007 to
2009. Catalogue Number: 82-623-X, Ottawa, Statistics
Canada.
[4] Kaczynski, A. and Henderson, K. (2008) Parks and re-
creation settings and active living: A review of associa-
tions with physical activity function and intensity. Jour-
nal of Physical Activity and Health, 5, 619-632.
[5] McNeill, L.H., Kreuter, M.W. and Subramanian, S.V.
(2006). Social environment and physical activity: A re-
view of concepts and evidence. Social Science & Me-
dicine, 63, 1011-1022.
doi:10.1016/j.socscimed.2006.03.012
[6] Booth, K.M., Pinkston, M.M. and Poston, W.S. (2005)
Obesity and the built environment. Journal of the Ame-
rican Dietetic Association, 105, S110-S117.
doi:10.1016/j.jada.2005.02.045
[7] McCormack, G., Giles-Corti, B., Lange, A., Smith, T.,
Martin, K. and Pikora, T.J. (2004) An update of recent
evidence of the relationship between objective and self-
report measures of the physical environment and physical
activity behaviours. Journal of Science and Medicine in
Sport, 7, 81-92. doi:10.1016/S1440-2440(04)80282-2
[8] Saelens, B.E. and Handy, S.L. (2008) Built environment
correlates of walking: A review. Medicine & Science in
Sports & Exercise, 40, S550-S566.
http://dx.doi.org/10.1249/MSS.0b013e31817c67a4
[9] Trost, S.G., Owen, N., Bauman, A.E., Sallis, J.F. and
Brown, W. (2002) Correlates of adults' participation in
physical activity: review and update. Medicine & Scie nce
in Sports & Exercise, 34, 1996-2001.
doi:10.1097/00005768-200212000-00020
[10] Feng, J., Glass, T.A., Curriero, F.C., Stewart, W.F. and
Schwartz, B.S. (2010) The built environment and obesity:
A systematic review of the epidemiologic evidence.
Health & Place, 16, 175-190.
doi:10.1016/j.healthplace.2009.09.008
[11] Powell, L.M., Slater, S., Chaloupka, F.J. and Harper, D.
(2006) Availability of physical activity-related facilities
and neighborhood demographic and socioeconomic cha-

S. A. Prince et al. / Open Journal of Preventive Medicine 1 (2011) 182-189
Copyright © 2011 SciRes. http://www.scirp.org/journal/OJPM/Openly accessible at
189
racteristics: A national study. American Journal of Public
Health, 96, 1676-1680. doi:10.2105/AJPH.2005.065573
[12] Gilliland, J., Holmes, M., Irwin, J.D. and Tucker, P.
(2006) Environmental equity is child’s play: mapping pu-
blic provision of recreation opportunities in urban neigh-
bourhoods. Vulnerable Children and Youth Studies, 1,
256-268. doi:10.1080/17450120600914522
[13] Riva, M., Gauvin, L. and Barnett, T. (2007) Toward the
next generation of research into small area effects on
health: a synthesis of multilevel investigations published
since July 1998. Journal of Epidemiol Community Health,
61, 853-861. doi:10.1136/jech.2006.050740
[14] Wilson, K., Eyles, J., Ellaway, A., Macintyre, S. and Ma-
cdonald, L. (2010) Health status and health behaviours in
neighbourhoods: A comparison of Glasgow, Scotland and
Hamilton, Canada. Health & Place, 16, 331-338.
doi:10.1016/j.healthplace.2009.11.001
[15] Parenteau, M.-P., Sawada, M., Kristjansson, E.A.,
Calhoun, M., Leclair, S., Labonté, R., et al. (2008) De-
velopment of neighborhoods to measure spatial indi-
cators of health. URISA Journal, 20, 43-55.
[16] North American Industry Classification System (NAICS).
Catalogue Number: 12-501-XIE, 2007, Ottawa, Statistics
Canada.
[17] Low income cut-offs. Catalogue no. 13-551-XIB. 1999.
Ottawa, Ontario, Statistics Canada.
[18] Heil, D.P. (2006) Predicting activity energy expenditure
using the Actical® activity monitor. Research Quarterly
for Exercise & Sport, 77, 64-80.
[19] Esliger, D.W., Probert, A., Connor-Gorber, S., Bryan, S.,
Laviolette, M. and Tremblay, M.S. (2007) Validity of the
Actical accelerometer step-count function. Medicine &
Science in Sports & Exercise, 39, 1200-1204.
doi:10.1249/mss.0b013e3804ec4e9
[20] Colley, R., Connor-Gorber, S. and Tremblay, M.S. (2010)
Quality control and data reduction procedures for accele-
rometry-derived measures of physical activity. Health
Reports, 21, 63-70.
[21] Tremblay, M.S., Warburton, D.E.R., Janssen, I., Paterson,
D.H., Latimer, A.E., Rhodes, R.E., et al. (2011) New
Canadian Physical Activity Guidelines. Applied Physiology,
Nutrition and Metabolism, 36, 36-46.
doi:10.1139/H11-009
[22] Tudor-Locke, C., Hatano, Y., Pangrazi, R.P. and Kang, M.
(2008) Revisiting “how many steps are enough?”. Me-
dicine & Science in Sports & Exercise, 40, S537-S543.
doi:10.1249/MSS.0b013e31817c7133
[23] Colley, R. C. and Tremblay, M. S. (2011). Moderate an-
dvigorous physical activity intensity cut-points for the
Actical accelerometer. Journal of Sports Sciences, 29,
783-789. doi:10.1080/02640414.2011.557744
[24] Wong, S., Colley, R. C., Connor-Gorber, S. and Tremblay,
M. S. (2011). Accelerometer sedentary activity thresholds
for adults. Journal of Physical Activity & Health, 8,
587-591.
[25] Canadian Guidelines for Body Weight Classification in
Adults. Catalogue Number: H49-179/2003E, 2003, Otta-
wa, Health Canada.
[26] Van Dyck, D., Deforche, B., Cardon, G. and De Bour-
deaudhuij, I. (2009) Neighbourhood walkability and its
particular importance for adults with a preference for
passive transport. Health & Place, 15, 496-504.
http://dx.doi.org/10.1016/j.healthplace.2008.08.010
[27] Martin, S.L., Kirkner, G.J., Mayo, K., Matthews, C.E.,
Larry, J. and Hebert, J.R. (2005) Urban, rural, and re-
gional variations in physical activity. The Journal of
Rural Health, 21, 239-244.
doi:10.1111/j. 1748-0361.2005.tb00089.x
[28] Lebel, A., Pampalon, R., Hamel, D. and Theriault, M.
(2009) The geography of overweight in Quebec: A mul-
tilevel perspective. Canadian Journal of Public Health,
100, 18-23.
[29] Parks, S.E., Housemann, R.A. and Brownson, R.C. (2003)
Differential correlates of physical activity in urban and
rural adults of various socioeconomic backgrounds in the
United States. Journal of Epidemiology and Community
Health, 57, 29-35. doi:10.1136/jech.57.1.29
[30] Van Dyck, D., Cardon, G., Deforche, B., Sallis, J.F.,
Owen, N. and De Bourdeaudhuij, I. (2010) Neighbor-
hood SES and walkability are related to physical activity
behavior in Belgian adults. Preventive Medicine, 50,
S74-S79. doi:10.1016/j.ypmed.2009.07.027
[31] Van Dyck, D., Cardon, G., Deforche, B., Owen, N., Sallis,
J.F. and de Bourdeaudhuij, I. (2010) Neighborhood walk-
ability and sedentary rime in Belgian adults. American
Journal of Preventive Medicine, 39, 25-32.
doi10.1016/j.amepre.2010.03.004