The food environment and dietary intake: Demonstrating a method for GIS-mapping and policy-relevant research

Aims: The aims of this paper are (1) to assess if perceptions of the food environment are associated with select dietary intake by neighborhood, and (2) to map neighborhoodspecific findings, demonstrating a method for policyrelevant research. Methods: Using pre-collected data from a Philadelphia, PA community health survey, we aggregated individual-level data (n=4,434 respondents) to neighborhoods (n=381 census tracts), adjusting for conceptually-relevant sociodemographic factors. We estimated Spearman correlations between multivariable adjusted food-environment perceptions (perceived produce availability, supermarket accessibility, grocery quality) and select dietary intake (reported fruit-and-vegetable and fast-food consumption), and mapped variables by neighborhood using geographic information systems (GIS). Results: Difficulty finding fruits and vegetables, having to travel outside of one's neighborhood to get to a supermarket, and poor grocery quality were each directly correlated with fast-food intake (rho=0.21, 0.34, 0.64 respectively; p values <0.001); and inversely correlated with fruit-and-vegetable intake (rho=-0.35, -0.54, -0.56 respectively; p values <0.001). Maps identified neighborhoods within the city with the worst food-environment perceptions and poorest dietary intakes. Conclusion: Negative perceptions of the food environment were strongly correlated with less-healthy eating in neighborhoods. Maps showed the geographic areas of greatest concern. Our findings demonstrate a method that might be used prospectively in public health for policy planning (e.g. to identify neighborhoods most in need), or retrospectively for policy assessment (e.g. to identify changes in neighborhoods after policy implementation).