Spatiotemporal trends in diabetes-related mortality by school district in the state of Michigan.

Objectives This study examined the spatiotemporal epidemiological status of diabetes-related death in relation to school district boundaries in the state of Michigan, USA. Methods A retrospective observational study was conducted using death records spanning the years 2007 - 2014 in Michigan with school district as the geographic unit of analysis. Geocoding was performed for each death record. Cluster analysis used spatial autocorrelation with local Moran's I, and spatiotemporal analysis used the Space Time Pattern Mining tool. Results The study revealed spatial clusters of high-high locations of diabetes-related mortality rate by school district in Michigan from 2007 - 2014. Spatiotemporal analysis showed the grids with intensifying, consecutive, sporadic and persistent hotspots of diabetes-related death at Lansing, Royal Oak, Flint City, Berkley, Detroit City, East Lansing, South Lake and Holt Public school districts. These school districts may be prioritized for school-based diabetes prevention programs. Conclusion The study demonstrated the presence of various hotspots of diabetes-related deaths within the Michigan State across the eight-year period of analysis. Understanding spatial and temporal hotspots could further improve our ability to evaluate diabetes burden across both time and space.