Introduction: The food environment has received considerable attention as a potential correlate of diet-related CV health. Features of the food environment are assumed to influence the places where people shop, which in turn affect the types of foods purchased and ultimately consumed. However, this assumption remains largely untested due to limited data on where people actually shop. Leveraging unique spatial data on food shopping destinations, we examine whether features of the food environment are associated with the food shopping practices of low-income African American women in Atlanta, Georgia (n=199). Hypothesis: Access to food retailers will be positively associated with their use. Methods: We collected spatial data on locations visited in the past month using a Google Maps-driven questionnaire. Food shopping practices, including the number of retailers, trip frequency, and average distance traveled from home, were assessed separately for supermarkets, convenience stores, and discount (e.g. dollar) stores. We used secondary data on retail food outlets to calculate three measures of food access: density, proximity, and quality. GEE estimated adjusted associations between food access and food shopping practices, controlling for correlation within census tracts. Adjusted models examined supermarket shopping practices, the only outlet used by a majority of women. Results: Only 12.1% of women had a supermarket located within 1 mile of home, compared to 53.8% for convenience and 22.6% for discount stores. Women lived farthest from supermarkets ( Mdn =1.6 mi, IQR =0.9-2.3), followed by discount ( Mdn =1.0 mi, IQR =0.6-1.3) and convenience ( Mdn =0.4 mi, IQR =0.2-0.7) stores. Women also traveled farthest to shop at supermarkets ( Mdn =3.6 mi, IQR =2.5-5.3), traveling over twice the distance of the closest outlet. Nearly all women (97%) shopped at supermarkets, 33% at discount stores, and 25% at convenience stores. In adjusted models, exposure to higher supermarket density was associated with a 13% decrease in distance traveled from home to utilized supermarkets ( e β =0.87, 95% CI : 0.80-0.94). Living farther away from a supermarket was weakly associated with distance traveled from home to utilized supermarkets ( e β =1.04, 95% CI : 1.03-1.06). Conclusions: Women may seek to overcome limited access to supermarkets, rather than rely on more accessible, but less healthy retailers. While women exposed to a higher density of supermarkets used supermarkets closer to home, living in closer proximity to a supermarket minimally affected how close to home they shopped. Higher retail density offers a greater number of supermarket choices, making it more likely that women will find one closer to home aligned with their personal preferences (e.g. affordability, hours, access from public transport). Interventions to improve food access must acknowledge that proximity alone may not drive outlet use.
Read full abstract