p - MD Anderson Cancer Center

Report
Neighborhood matters:
How characteristics of the
residential environment relate
to physical activity, sedentary
behavior, and body mass index
among African American
adults
Larkin L. Strong, PhD
Lorraine R. Reitzel, PhD
University of Texas MD Anderson Cancer Center
Department of Health Disparities Research
Agenda
• Brief overview and study sample
• A tale of two studies
– Study 1: Associations of perceived neighborhood
physical and social environments with physical
activity and television viewing in African American
men and women
– Study 2: Density and proximity of fast food
restaurants and body mass index among African
Americans
Brief Overview
• Growing interest in how neighborhood factors
influence health and health behaviors over and
above individual-level factors
• Very little research has focused on African
American populations in this regard
• African Americans at particular risk of health
disparities
• More understanding is needed to inform policy
and interventions to affect these disparities
Study Sample
• Project CHURCH
– Creating a Higher Understanding of cancer Research and
Community Health
– Designed to assess behavioral, social, and environmental
cancer risk factors in 1,501 African American adults
– Cross-sectional analysis of self-reported baseline data
collected in 2008-2009
• Setting
– Large mega-church in Houston, TX
• Sampling Protocols
– Participants recruited through church media channels
– Inclusion: >age 18; residence in Houston area; functional
telephone number; must attend church
Study 1
Strong LL, Reitzel LR, Wetter DW, McNeill LH
American Journal of Health Promotion.
2013;27(6):401-9.
Introduction
• Most adults in the US
– Are not physically
active
– Spend over 50% of
their waking time in
sedentary behaviors
• Some racial/ethnic
groups are
disproportionately
affected
60
Moderate-to-Vigorous
Physical Activity (PA)
40
%
20
0
CDC BRFSS, 2011
TV Viewing
8
6
Hrs/
day
4
2
0
White
The Nielsen Company, 2011
Black
Hispanic
Asian/other
The Role of the Environment
• Strong evidence that
characteristics of the
built environment are
associated with PA
• What about sedentary
behavior?
• What about the role of
the neighborhood social
environment?
Study Purpose
• To investigate the associations of perceived
aspects of neighborhood social and physical
environments with PA and TV viewing in a large,
church-based sample of African American men
and women
Methodology (cont’d)
• Measures
– Outcomes
• Meeting PA guidelines (International Physical Activity
Questionnaire; yes/no)
• Average TV viewing time per day (log-transformed)
– Neighborhood Perceptions
• Social Cohesion and Trust (Sampson et al., 1997)
• Neighborhood Problems (Steptoe & Feldman, 2001)
• Analysis
– Multivariate generalized estimating equations
• Stratified by gender
• Controlled for sociodemographics
Results: Participant Characteristics
N= 1,374
–
–
–
–
–
–
–
Age = 45.5 (12.6)
76% female
50% <Bachelor’s degree
45% married/living with partner
25% unemployed
25% <$40,000 annual income
BMI = 31.7 (7.2)
•
29.8% overweight, 53.9% obese
Results: Neighborhood Problems
Percent reporting neighborhood conditions as a problema
Men
(n=349)
%
Women
(n=1,025)
%
Litter in the streets
30
33
Smells and fumes
Walking around after dark**
14
27
16
40
Problems with dogs**
Noise from traffic or other homes*
Lack of entertainment
Traffic and road safety problems
24
22
22
21
34
29
24
22
Lack of places to shop
Vandalism*
22
33
25
39
Disturbances by neighbors/youngsters
29
26
aConditions
dichotomized to represent “some problem/serious problem” and “not a problem”
Results: Associations between
Neighborhood Perceptions and Behaviors
Physical Activity
Log-transformed TV Viewing
Men
Women
Men
Women
OR (95% CI)
OR (95% CI)
β (SE)
β (SE)
Neighborhood
Problems
0.92 (0.84, 1.00)
0.99 (0.95, 1.03)
.002 (0.010)
.017* (0.006)
Social
Cohesion
1.07 (0.98, 1.17)
1.06* (1.02, 1.22)
.005 (0.011)
-.014 (0.008)
Scales
Models adjusted for sociodemographics (age, education, income, employment, marital status,
presence of children in home, number of years in neighborhood)
*p<0.01
Predicted Outcomes
Hrs/day
Predicted Mean Daily TV Viewing
5
4.5
4
3.5
3
2.5
2
1.5
1
0.5
0
Predicted Probabilities of
Meeting PA Guidelines
1
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
10
20
30
Neighborhood Problems
Women
Men
5
10
15
20
Social Cohesion
Women
Men
25
Results: Associations between Specific
Conditions and Behaviors
Physical Activity
Log-transformed TV Viewing
Men
Women
Men
Women
OR (95% CI)
OR (95% CI)
β (SE)
β (SE)
Litter in the street
0.66 (0.33, 1.32)
1.30 (0.97, 1.73)
-0.064 (0.076)
0.115* (0.048)
Walking around
after dark
0.48* (0.24, 0.98)
0.81 (0.64, 1.03)
-0.014 (0.075)
0.110* (0.043)
Traffic and road
safety problems
0.37** (0.18, 0.73)
0.90 (0.64, 1.27)
0.098 (0.082)
0.022 (0.049)
0.49 (0.24, 1.02)
0.78 (0.57, 1.09)
Individual Conditionsa
Lack of places to
shop
-0.002
(0.074)
0.103* (0.043)
Models adjusted for sociodemographics (age, education, income, employment, marital status, presence of
children in home, number of years in neighborhood)
aConditions dichotomized to represent “some problem/serious problem” and “not a problem”
*p<0.05
**p<0.01
Limitations
• Cross-sectional data
• Self-report
– Recall, social desirability bias, e.g. IPAQ
• Neighborhood data are subjective, may not
represent actual conditions
• Convenience sample of church-based African
American adults
– May not be generalizable
Conclusions
• Among the first studies to suggest that social and
physical aspects of neighborhood environments
may affect sedentary in addition to active
behaviors
• Social cohesion was positively associated with
PA, although only significant in women
• Perceiving greater disorder within neighborhood
was associated with increased TV viewing in
women
• Identified specific neighborhood conditions
associated with PA in men and TV viewing in
women
Implications
• Important to consider neighborhood social
characteristics and the design and conditions of
physical environment for intervention/policy
efforts
• Intervention strategy – facilitate positive
interactions among residents while also
promoting healthy behaviors
• Additional research is needed to understand the
mechanisms through which neighborhood
attributes affect behavior
Study 2
Reitzel LR, Regan SD, Nguyen N, Cromley EK,
Strong LL, Wetter DW, McNeill LH
American Journal of Public Health. Published
online ahead of print May 16, 2013: e1-e7.
doi:10.2105/AJPH.2012.301140
Background
• Racial/ethnic disparities in obesity prevalence
– 38.8% African American vs 36.2% White men
– 58.5% African American vs 32.2% White women
• BMI gap is widening too
• African American neighborhoods have higher
density of FFRs than White neighborhoods
• Studies suggest FFR availability and FF
consumption stronger among non-Whites than
Whites
• African Americans may be more likely to consume
FF if available, and it is more likely to be available
The Current Study
• Recent study examined FFR density and BMI
among 4500 African Americans from Jackson
– Null relations
• Other studies in area, mixed results overall
• No previous studies have looked at FFR
proximity among African Americans and
relations with BMI
• No previous studies have examined income as a
moderator, despite that reasons for frequent FF
consumption = accessibility and affordability
• Current study was meant to redress these gaps
Methodology
• Obtained FFR addresses from InfoUSA
• FFRs = limited service restaurants, hamburger/
hotdog establishments
• Geocoded residential addresses of study
sample and FFR addresses
• Calculated FFR density at 0.5, 1, 2, and 5 miles
around participant’s homes
• Calculated proximity to closest FFR from the
home
• BMI – calculated from height/weight measured
twice by stadiometer/scale
Fast Food Restaurants
Variable of Interest - Density
Fast Food Restaurants
1 mile Density
Buffer*
Participant’s Home
* Note: Density Buffer s were calculated at .05, 1, 2, and 5 miles for this study. 1 mile buffer presented as an example.
Variable of Interest - Proximity
Fast Food Restaurants
Proximity to Closest
Fast Food Restaurant
Participant’s Home
Statistical Approach
• Adjusted generalized linear regression models,
without and later with interaction term (income)
• Analyses controlled for:
– sociodemographics: age, gender, partner status,
total annual household income, educational level,
and employment status
– tenure in years at the reported home address
– presence of children in the home
– physical activity level
– television viewing time
– neighborhood median household income
Results: FFR Density & BMI
• Mean # of FFRs
– 2.5 (1.9) in .5 mile buffer; 4.5 (4.2) in 1 mile
buffer; 11.4 (9.8) in 2 mile buffer; 71.3 (50.4) in 5
mile buffer
• Main effects were non-significant
• Significant interaction terms at 0.5, 1, & 2 miles
• Stratified analyses: density = BMI for
participants earning <$40,000/year
– 0.5 mile buffer (b = 1.15; p= .009)
– 1 mile buffer (b = 1.23; p= .008)
– 2 mile buffer (b = 1.69; p= .025)
Figure 1:
Adjusted
relations of
FFR density
within a .5
mile buffer
and predicted
BMI by
income
higher density values = greater density
Results: FFR Proximity & BMI
• Mean FFR Proximity = 1.01 miles (.77)
• Lower income participants were more likely to live
closer to FFR than higher income participants
• Main effects: closer proximity = BMI (b= 0.98; p<
.001)
• Significant interaction term (p = .029)
• Stratified analyses: closer proximity = BMI for
both income groups
– <$40,000/year (b= 0.92; p= .013)
– >$40,000/year (b= 0.99; p= .014)
• Every mile closer to a FFR = 2.4% higher BMI
Figure 1:
Adjusted
relations of
FFR proximity
and predicted
BMI by
income
higher proximity values = greater distance
Implications
• Why links between FFR density and BMI among
lower income participants?
– More affordable? More convenient? More
exposure (cueing)? Places for socialization?
Transportation issues?
• Why links between FFR proximity and BMI?
– Only need 1 to purchase FF? Ease and
convenience? Cueing? Transportation?
• Utility of zoning laws or conditional use permits
to regulate locations and numbers of FFRs
around residential areas
Limitations
• Cross-sectional
• Need to assess consumption frequency and
consumption choices
• Need to understand entire food landscape
• No information on disposable income
• No information in FFR locations around other
locations of importance (e.g., work)
• Sample was mostly female, well-educated,
church-going, metropolitan/urban
Research Support
– The University Cancer Foundation
– The Duncan Family Institute
– The Ms. Regina J. Rogers Gift: Health Disparities
Research Program
– The Cullen Trust for Health Care Endowed Chair
Funds for Health Disparities Research
– The Morgan Foundation Funds for Health
Disparities Research and Educational Programs
– The National Cancer Institute through The
University of Texas MD Anderson's Cancer
Center Support Grant (grant number CA016672)

similar documents