One of the greatest public health challenges in the US is the epidemic of obesity. Support for an association between the environment and physical activity, eating behavior and, to a lesser extent, obesity has been recently documented. While these associations may represent a causal relationship, there are several possible alternative explanations that previous research has not adequately ruled out. For example, a person that values and enjoys walking is more likely to move to a walkable neighborhood;so the effect of neighborhood is likely confounded with individual attributes. The Moving to Opportunity for Fair Housing Demonstration (MTO) provides an excellent setting to examine the role of neighborhood characteristics on obesity, physical activity and fruit/vegetable intake. The goal of MTO was to determine whether persons who moved to "better" neighborhoods would experience improvements in a variety of outcomes such as education, economic and health compared to similar persons who did not move. To ensure that comparisons were made between similar persons, families were randomly assigned to one of three groups: the experimental group (who were required to move to census tracts where less than 10% of persons lived in poverty), the Section 8 group and the control group. The key advantage of the MTO data is participants were randomized to treatment and control groups, which would likely eliminate any baseline differences among these groups. Since many weight loss interventions fail, an important finding of the MTO study was that adult participants (N=3526) in the experimental group had a significantly lower BMI 4 to 7 years after randomization compared to controls. Additionally, those in the experimental group had a higher frequency of fruit and vegetable consumption and those in the Section 8 experimental group had a greater frequency of physical activity relative to controls. Thus the purpose of our research is to examine whether the differences in these outcomes seen across study groups were mediated by attributes of the participant's environment. In other words, we will examine whether those in the experimental groups were more likely to be exposed to supportive environments, and whether differences in environmental characteristics explain the treatment (i.e., randomization) effect. This study will link MTO Interim Evaluation data with environmental characteristics of participant's neighborhoods. The MTO Interim Evaluation included information on health outcomes including BMI, physical activity levels and consumption of fruit and vegetables along with demographic information. Environmental characteristics that will be collected and linked with the Interim Evaluation include: food availability, food prices, commercially and publicly available physical activity resources, city-sponsored recreation centers, connectivity of streets, land use, residential density, crime statistics and census data.
The relevance of this research to public health is threefold. It will provide perhaps the best evidence to date on whether environmental attributes are causally related to obesity levels, physical activity and fruit and vegetable consumption at the individual level;it will provide information to inform policy makers on how to better spend public housing financial resources to reduce the burden of obesity in this population;and it will suggest whether built environment changes are a cost effective strategy to reduce the epidemic of obesity.