Evaluation of Mediation via the Physical Activity Environment of Neighborhood, Socioeconomic Environment, and Type 2 Diabetes Associations

Inequitable access to neighborhood leisure-time physical activity (LTPA) resources may explain geographic disparities in type 2 diabetes (T2D) . We evaluated the LTPA environment as a potential mediating pathway through which the neighborhood socioeconomic environment (NSEE) affects T2D. We used harmonized measures and analytic approaches in three independent study samples in the Diabetes Location, Environmental Attributes, and Disparities Network: the national Veterans Administration Diabetes Risk (VADR) cohort, comprising electronic health records (EHR) of 4.1 million T2D-free veterans; Reasons for Geographic and Racial Differences in Stroke (REGARDS) , a national prospective cohort (n=11,2T2D-free) ; and a case-control study in the Geisinger EHR in Pennsylvania (15,888 T2D cases) . We measured NSEE with an index of census tract-level Census metrics (higher indicates disadvantage) . The LTPA environment was evaluated by PA facility density within street network buffers and population-weighted distance to parks. In models stratified by levels of community urbanicity, we estimated natural direct effects and indirect effects for each mediator. PA facility density mediated NSEE-T2D effects, but the magnitude and direction of the indirect effects differed by urbanicity and study sample. In rural communities, we found positive indirect effects via PA facility density (incidence rate difference [95% confidence interval] of NSEE quartile 4 vs. 1: 0.0[0.0038, 0.0099] in VADR; 1.53 [0.25, 3.05] in REGARDS) , representing a proportion mediated of 3% (VADR) and 51% (REGARDS) . We found small indirect effects for all other community types in VADR but none in the Geisinger sample. Distance to parks mediated NSEE-T2D effects only in VADR rural communities. Results suggest regional or population-specific differences in how and whether the LTPA environment mediates NSEE-T2D relations, which could support local T2D prevention efforts. Disclosure K.A. Moon: None. C.M. Nordberg: None. A. Zhu: Research Support; Amgen Inc. J. Uddin: None. P.M. Lopez: None. S.L. Orstad: None. M.D. Schwartz: None. V. Ryan: None. B.S. Schwartz: None. A.P. Carson: Research Support; Amgen Inc. D. Long: Research Support; Amgen Inc. J.R. Brown: None. G.S. Lovasi: None. S. Adhikari: Research Support; Johnson & Johnson. R. Kanchi: None. S. Avramovic: None. K.R. Siegel: None. M.N. Poulsen: None. Funding Centers for Disease Control (U01DP006293, Drexel University; U01DP006296, Geisinger-Johns Hopkins University; U01DP006299, New York University School of Medicine; U01DP006302, University of Alabama at Birmingham) .