Spatial Pattern Analysis of Diabetes Rates Nationwide as a Function of Demographic Variables

The Centers for Disease Control and Prevention (CDCP) estimates that diabetes affects 25.8 million people in the U.S. Diabetes is a group of diseases marked by high levels of blood glucose resulting from defects in insulin production, insulin action, or both.  Diabetes was the seventh leading cause of death based on U.S. death certificates in 2007, with total  costs of the disease estimated at $174 billion. County-level estimates of diagnosed diabetes seem to indicate clustered patterns of in areas of the southern U.S. and other spatially correlated relationships.   This presentation will report on a neighborhood analysis to compare diagnosed diabetes patterns with demographic and economic indicators, specifically related to income.  Our goal is to use ordinary least squares (OLS) and geographically weighted regression (GWR) to identify relationships between diagnosed diabetes and income-related characteristics derived from the U.S. Census. Including the health disparities data that show  there are income and ethnic backgrounds shown  typically to be co-factors in lower health status.  Our hypothesis is that diagnosed diabetes rates are not only correlated with income, but are also geographically weighted in this correlation.  Our method is to use geographic information systems (GIS) software via spatial analyst tools to perform GWR using data obtained from the CDCP and U.S. Census.  This project is significant in that it will aid in determining whether spatial patterns of diabetes are also correlated with income patterns, and whether broad efforts to stem this disease can be targeted to particularly active zones where the disease is most prevalent.