Obesity prevalence increased and F/V consumption decreased with increasing distance to supermarket in metropolitan areas, but not in nonmetropolitan areas. These results suggest that there may be a threshold distance in nonmetropolitan areas beyond which distance to supermarket no longer impacts obesity and F/V consumption. In addition, obesity and food environments in nonmetropolitan areas are likely driven by a more complex set of social, cultural, and physical factors than a single measure of supermarket accessibility. Future research should attempt to more precisely quantify the availability and affordability of foods in nonmetropolitan areas and consider alternative sources of healthy foods besides supermarkets.
Distance to supermarket was greater in nonmetropolitan than in metropolitan areas. The odds of obesity increased and odds of consuming F/V five times or more per day decreased as distance to supermarket increased in metropolitan areas for most store size categories. In nonmetropolitan areas, however, distance to supermarket had no associations with obesity or F/V consumption for all supermarket size categories.
Limited access to supermarkets may reduce consumption of healthy foods, resulting in poor nutrition and increased prevalence of obesity. Most studies have focused on accessibility of supermarkets in specific urban settings or localized rural communities. Less is known, however, about how supermarket accessibility is associated with obesity and healthy diet at the national level and how these associations differ in urban versus rural settings. We analyzed data on obesity and fruit and vegetable (F/V) consumption from the Behavioral Risk Factor Surveillance System for 2000-2006 at the county level. We used 2006 Census Zip Code Business Patterns data to compute population-weighted mean distance to supermarket at the county level for different sizes of supermarket. Multilevel logistic regression models were developed to test whether population-weighted mean distance to supermarket was associated with both obesity and F/V consumption and to determine whether these relationships varied for urban (metropolitan) versus rural (nonmetropolitan) areas.