Spatial Analysis and Correlates of County-Level Diabetes Prevalence, 2009 - 2010
ORIGINAL RESEARCH — Volume 12 — May 7, 2015
The figure consists of 4 maps of the contiguous US states with estimates and t values of the variables poverty and nonwhite based on a geographically weighted regression (GWR) for county-level diabetes prevalence. Map A shows the estimates of poverty ranging from −0.03 to 0.14. The variable represents the percentage of population living below the federal poverty level in a county. Based on map B with t values for the estimates of poverty, the negative effect sizes are not significant. Of the positive estimates, 40% are significant. The greatest effect, with significance, of poverty on diabetes is in the Southwest and Southern California, southern part of Florida, and in the Northeast.
Map C displays the GWR estimates of nonwhite population ranging from −0.04 to 0.09. This variable represents the percentage of population that identifies as nonwhite in a county regressed on diabetes prevalence. Based on map D, 78% of the county estimates are significant. The greatest effect sizes for the nonwhite population on diabetes prevalence are in the Dakotas south through Nebraska and into northern Kansas and Missouri.
Figure 1. Spatial variation in parameter estimates and t values in US counties for the percentage of people living below the federal poverty level (maps A and B) and the percentage of nonwhite population (maps C and D). Data sources: American Community Survey (2006–2010) (13) and Centers for Disease Control and Prevention (12).
The figure consists of 5 maps of the 48 contiguous US states. Maps A through D have estimates and t values for the variables walking or biking to work and physical inactivity based on a geographically weighted regression (GWR) for county-level diabetes prevalence. Map A shows the estimates of the variable walking or biking to work ranging from −32.12 to 6.70 per county. This variable represents the percentage of people in a county who walked or biked to work in the preceding week. Only 30% of the estimates are significant, which is shown in map B with t values. There are 2 main regions where walking or biking to work has a significant inverse relationship with diabetes prevalence (higher rate of walking or biking to work is associated with lower diabetes prevalence). One region follows the spine of the Rocky Mountains. The second region is in the Midwest and the Mississippi and Ohio River valleys. There are 5 counties where walking and biking has a significant positive relationship with diabetes prevalence, each in South Dakota.
Map C shows estimates for physical inactivity ranging from −0.08 to 0.11. This variable represents the percentage of people in a county who did not participate in physical activity or exercise in the preceding 30 days, regressed on diabetes prevalence. Only 18% of the county estimates for this variable are significant based on the t values in map D. Physical inactivity has a significant inverse relationship with diabetes in the Northwest. It has significant positive relationship (higher rates of physical inactivity associated with higher rates of diabetes) in New Mexico and parts of Texas.
Map E displays the local R-squared values based on the GWR for diabetes prevalence. Local R-squared values range from 0.06 to 0.94. Local R-squared represents the amount of variance in diabetes prevalence explained by the model in each county. The model performed the best in parts of the country stretching from the Dakotas and Montana south to New Mexico and Arizona, as well as in the Southeast. The GWR model was not able to explain much variance in the Northwest, parts of the upper Midwest, and the Mid-Atlantic.
Figure 2. Spatial variation in parameter estimates and t values in US counties for percentage of employed population walking or cycling to work (maps A and B) and the percentage of the population that is physically inactive (maps C and D); local R-squared for full geographically weighted regression model (Map E). Data sources: American Community Survey (2006–2010) (13) and Centers for Disease Control and Prevention (12).
The opinions expressed by authors contributing to this journal do not necessarily reflect the opinions of the U.S. Department of Health and Human Services, the Public Health Service, the Centers for Disease Control and Prevention, or the authors’ affiliated institutions.