Development of a national childhood obesogenic environment index in the United States: differences by region and rurality

Background: Diverse environmental factors are associated with physical activity (PA) and healthy eating (HE) among youth. However, no study has created a comprehensive obesogenic environment index for children that can be applied at a large geographic scale. The purpose of this study was to describe the development of a childhood obesogenic environment index (COEI) at the county level across the United States. Methods: A comprehensive search of review articles (n = 20) and input from experts (n = 12) were used to identify community-level variables associated with youth PA, HE, or overweight/obesity for potential inclusion in the index. Based on strength of associations in the literature, expert ratings, expertise of team members, and data source availability, 10 key variables were identified - six related to HE (# per 1000 residents for grocery/superstores, farmers markets, fast food restaurants, full-service restaurants, and convenience stores; as well as percentage of births at baby (breastfeeding)-friendly facilities) and four related to PA (percentage of population living close to exercise opportunities, percentage of population < 1 mile from a school, a composite walkability index, and number of violent crimes per 1000 residents). Data for each variable for all counties in the U.S. (n = 3142) were collected from publicly available sources. For each variable, all counties were ranked and assigned percentiles ranging from 0 to 100. Positive environmental variables (e.g., grocery stores, exercise opportunities) were reverse scored such that higher values for all variables indicated a more obesogenic environment. Finally, for each county, a total obesogenic environment index score was generated by calculating the average percentile for all 10 variables. Results: The average COEI percentile ranged from 24.5-81.0 (M = 50.02, s.d. = 9.01) across US counties and was depicted spatially on a choropleth map. Obesogenic counties were more prevalent (F = 130.43, p < .0001) in the South region of the U.S. (M = 53.0, s.d. = 8.3) compared to the Northeast (M = 43.2, s.d. = 6.9), Midwest (M = 48.1, s.d. = 8.5), and West (M = 48.4, s.d. = 9.8). When examined by rurality, there were also significant differences (F = 175.86, p < .0001) between metropolitan (M = 46.5, s.d. = 8.4), micropolitan (M = 50.3, s.d. = 8.1), and rural counties (M = 52.9, s.d. = 8.8) across the U.S. Conclusion: The COEI can be applied to benchmark obesogenic environments and identify geographic disparities and intervention targets. Future research can examine associations with obesity and other health outcomes.