1649-P: Community Predictors of Diabetes Prevalence and Change in Burden Over Time: U.S. Counties, 2003-2012

Abstract

Introduction: This study applies a novel epidemiologic approach to perform temporal, multilevel analysis of nationally representative, repeated cross-sectional data to assess the relationship between diabetes and various county-level environmental measures, including socioeconomic context, walkability, and the food environment. Methods: We obtained data from the 2003-2012 Behavioral Risk Factor Surveillance System, an ongoing, nationally representative, cross-sectional telephone survey of the U.S. noninstitutionalized population ages ≥ 18 years. County-level variables included economic (unemployment), food (convenience store and fast food density), and built environment (active commuting) measures. Using random-effects within-between (REWB), we fit a logistic mixed-effects model to estimate associations between community exposures and odds of prevalent diabetes, adjusted for individual-level demographic and other characteristics. REWB models incorporate temporality into analyses of repeated cross-sectional surveys by using group-mean-centering. Results: Age-adjusted diabetes prevalence across the study period was 7.9%. County unemployment, fast food density, and convenience store density were positively associated with diabetes prevalence, and active commuting had an inverse association. Odds ratios ranged from 0.95 for an increment in active commuting to 1.05 for convenience store density. Within-county associations were attenuated for convenience stores and fast food restaurants, but within-county change in active commuting was higher (aOR: 0.92; 95% CI: 0.82, 1.02). Discussion: These analyses demonstrate the utility of REWB models in evaluating effects of group-level exposures on diabetes risk using repeated cross-sectional survey data and provide preliminary evidence towards our hypotheses of community-level drivers of diabetes risk. Disclosure P.M. Lopez: None. J. Feldman: None. P.E. Rummo: None. B. Elbel: None. L. Thorpe: None.