Association between Neighborhood Food Access, Household Income, and Purchase of Snacks and Beverages in the United States.

Considerable research on the risk factors of obesity and chronic diseases has focused on relationships between where people live, where they shop, and the types of food they purchase. Rarely have investigators used a national sample and explicitly addressed the amount of energy-dense and nutrient-poor foods purchased in different types of neighborhood food stores. Even more rarely have studies accounted for the characteristics of the broader built environment in which food stores are located and which affect the convenience of using neighborhood food stores. We used a large population-based cohort of predominantly white U.S. households from the Nielsen Homescan Consumer Panel 2010 dataset to examine whether there were positive cross-sectional associations between availability of neighborhood convenience stores and supermarkets and self-reported household annual expenditures for snacks and beverages. We examined this relationship separately for poor and non-poor households as defined by the 2010 U.S. federal poverty threshold. We used mixed error-component regression models to examine associations between availability of neighborhood food stores and the expenditures on snacks and beverages, controlling for regional destination accessibility, availability and diversity of neighborhood destinations, and neighborhood street connectivity. In multivariate analyses, we observed that poor households in neighborhoods with few convenience stores purchased more snacks than poor households in neighborhoods with many convenience stores (b = -0.008, < 0.05). Non-poor households in neighborhoods with many convenience stores and fewer supermarkets purchased more snacks than non-poor households in neighborhoods with few convenience stores and many supermarkets (b = 0.002, < 0.05 for convenience stores; b = -0.027, < 0.05 for supermarkets). Increase in number of convenience stores decreased the purchase of snacks by poor households, but increased in non-poor households. On other hand, increase in number of supermarkets discouraged purchase of snacks by non-poor households but had no effect on the purchasing behavior of the poor-households.Therefore, evaluation of access to energy-dense and nutrient-poor foods should include a consideration of geographic proximity. Local governments should consider strategies to expand the availability and access to nutrient-rich food and beverage products in convenience stores for consumers.