1. Exploring the Relationship between Obesity and the Built Environment through Urbanicity
of U.S. Counties
Sociology Honors: Olivia Godsil, Mentor: Sara Curran
Introduction
• Question: Is there an association between obesity and urban status
at the county level in the United States?
• Hypothesis:
• Obesity is a significant public health concern in the industrialized
world: an epidemic resulting in fatal but preventable diseases (CDC
2014; WHO 2015)
• The proportion of the world population
living in cities is increasing (UN 2014)
• Urban status = the size of the city
(population density) in combination with
how people behave within that built
environment (infrastructure)
• Social determinants of health = the conditions of the social
environment in which one lives and grows
• Social determinants of health view: changes in a person’s everyday
encounters with the surrounding infrastructure can better prevent or
lessen obesity
• Genetics and behaviors do not explain it all
Methods
• Merged four data sets from the CDC (2010), the Census (2000), and
two American Community Surveys 5-year average (2006-2010)
• County level (3,076 counties)
• Dependent variable: obesity prevalence (measured in 2010)
• BMI measure = weight divided by height squared (kg/m2)
• Independent variables (measured prior to 2010):
• Urban status (urban-rural classification)
• Population density (individuals per square mile)
• Average commute times (in minutes)
• Commute methods (walk, bike, public transit, drive, home)
• Controls: Age-adjusted obesity prevalence and sex
• Step 1: Bivariate analysis - Scatter plot of population density and
obesity prevalence
• Step 2: OLS multiple regression of population density, average
commute times, and modal commute methods on obesity prevalence
• Step 3: Estimates of the impact of urban status on obesity prevalence
based on equation results from step 2
Discussion
• Commuting behavior mediates the relationship between obesity and population density
• As expected, areas with shorter commute times and more individuals walking, biking, or riding transit to work tend to have
fewer obese residents
• The importance is not in urban status so much as the type of city one lives in
• Cities with better infrastructures tend to provide more opportunities for physical activity
• Cities could really be the first place to make a difference since they are becoming the primary place of residence
• Manageable enterprising actions policy wise: strategic urban planning and city
planning that encourages walking, biking, and transit use
• Outside of cities residents are spread out, which makes policy harder to enforce
• Where an individual lives matters!
• Walkability is important
Urban
Status
Obesity
Results
Table 2. Impact Assessment of Urban Status on Obesity Prevalence
Better than Average Public Transit
(Mean +2 s.d.)
Better than Average Walking and Biking
(Mean +2 s.d.)
Expected Change in Obesity Prevalence - 5.9% - 3.58%
Acknowledgements
A special thank you to Hedwig Lee, Tim Thomas, Frank Edwards, and Patty Glynn for all of the guidance and support!
Marshall, Sean. 2012. “Striving for Transit-Friendly Communities in the Puget Sound Region.” Smart Growth America.
Parkford, Stan. 2014. “Polk Street Contra-Flow Bike lane Opens to the Public.” Streets Blog SF.
ObesityCity
Daily Routine
Table 1. Relationships between Obesity and Select Social
Determinants of Health. Adults in U.S. Counties, 5-year average
(2006-2010).
Estimate (Standard
Error)
Population Density 3.074*** (0.728)
Percent Female 0.012 (0.034)
Average Commute Time (in minutes) 0.019*** (0.005)
Percent who Walk or Bike to work -0.435*** (0.068)
Percent who take Public Transit -0.822*** (0.070)
Percent who Drive -0.341*** (0.057)
Percent who work from Home -0.638*** (0.058)
Constant 65.376*** (5.891)
R2
0.162 (4.066)
Note: *p<0.05; **p<0.01; ***p<0.001; (two-tailed tests).
Table 1. Relationships between Obesity and Select Social
Determinants of Health. Adults in U.S. Counties, 5-year average
(2006-2010).
Social Determinant of Health Estimate (Standard Error)
Population Density -1.518*** (0.126)
Percent Female 0.128*** (0.035)
Average Commute Time (in minutes) 0.018*** (0.005)
Percent who Walk or Bike to work -0.405*** (0.066)
Percent who take Public Transit -0.452*** (0.062)
Percent who Drive -0.246*** (0.057)
Percent who work from Home -0.645*** (0.057)
Constant 53.285*** (5.843)
R2 0.195 (3.986)
Note: *p<0.05; **p<0.01; ***p<0.001; (two-tailed tests).