Decomposing Racial Disparities in Obesity Prevalence Variations in Retail Food Environment

Introduction: Racial disparities in obesity exist at the individual and community levels. Retail food environment has been hypothesized to be associated with racial disparities in obesity prevalence. This study aimed to quantify how much food environment measures explain racial disparities in obesity at the county level. Methods: Data from 2009 to 2010 on 3,135 U.S. counties were extracted from the U.S. Department of Agriculture Food Environment Atlas and the Behavioral Risk Factor Surveillance System and analyzed in 2013. Oaxaca-Blinder decomposition was used to quantify the portion of the gap in adult obesity prevalence observed between counties with a high and low proportion of African-American residents is explained by food environment measures (e.g., proximity to grocery stores, per capita fast-food restaurants). Counties were considered to have a high African-American population if the percentage of African-American residents was >13.1%, which represents the 2010 U.S. Census national estimate of percentage African-American citizens. Results: There were 665 counties (21%) classified as a high African-American county. The total gap in mean adult obesity prevalence between high and low African-American counties was found to be 3.35 percentage points (32.98% vs 29.63%). Retail food environment measures explained 13.81% of the gap in mean age-adjusted adult obesity prevalence. Conclusions: Retail food environment explains a proportion of the gap in adult obesity prevalence observed between counties with a high proportion of African-American residents and counties with a low proportion of African-American residents. (C) 2016 American Journal of Preventive Medicine