Innovative technology for the planning and evaluation of local community health: GIS modeling of diabetes and predictor variables

Efforts to innovatively improve and evaluate health conditions of communities, like those of the Centers for Disease Control and Prevention (CDC) Community Transformation Grants (CTG), face the burden of identifying substantive confluences of health inequities and chronic disease across space. Broward Regional Health Planning Council's “Transforming Our Community's Health” (TOUCH), made possible with funding from CDC CTG, employs GIS methods to locate areas of significant interaction between Long Term Diabetes rates (LTD) and actionable predictor variables for intervention and evaluation activities aimed at increasing healthy food consumption. To identify areas of need, monitor efforts to ameliorate disparities, and evaluate interventions, TOUCH correlated ZIP Code hospital discharge LTD with rates of African Americans (AA), Earned Income Tax Credit returns (EITC), and the proportion of unhealthy food retail to healthy retail square footage per population (FOOD). TOUCH analyzed all variables with GIS for autocorrelation (hot spots/cold spots) and employed spatial regression models to test the relationship between LTD and predictor variables. LTD positively correlates with rates of AA (+0.80), EITC (+0.82), and FOOD (+0.54). Controlling for autocorrelation, one regression model (R^2 =77) finds a significant association between LTD with AA (p<0.00), Unhealthy Food Retail (p<.0.00), and Healthy Food Retail (p<0.04). Based on the spatial distribution of LTD and its social and environmental determinants across 439 square miles, TOUCH established strategies for increasing healthy food consumption in targeted communities. These activities demonstrate how policymakers, researchers, and the public may employ GIS as an innovative technology to develop geo-strategically informed intervention and evaluation plans.