Is it the built environment? Assessment of food deserts, food venue choice, and physical actvity in the Northeast

The built environment, from the accessibility of public transportation to the geographic location of food purveyors may play an important role in contributing to obesity. Both of these and individual eating behaviors have changed over the past decades. Using a survey of residents of Vermont, Maine, and New Hampshire, this study explores the relationship of the built environment and individual food and activity choices and their contribution to obesity. Using GIS, and the statistical techniques of cluster analysis and recursive modeling of limited dependent variables, we estimate the impact of the built environment on food venue choices and obesity. Data are from a Department of Transportation funded (DOT) survey of the Northeast region of the U.S. The analysis uses 717 complete responses; 72% reside in rural areas; 42% are female. These data are combined with Nelsen Business point data and GIS analysis to identify the location of a variety of food venues, from convenience stores to full service restaurants. Results show, given the proximate nature of the Northeast, not even the most rural respondent could be considered to be living in a food desert. Cluster analysis revealed three food venue patterns: at home, away from home, and mixed. Using a two stage estimation procedure,and controlling for various individual and environmental characteristics, food venue choice does not impact obesity when type of eating pattern and physical activity patterns are accounted for. In terms of public policy, single, nationwide definitions of food deserts may not be appropriate. In addition, evidence points toward wider restaurant menu labeling policies, and further examination of nutrition labeling policies.