Objective

To examine differences in the incidence and prevalence of diagnosed diabetes by county rurality.

Patients and methods

This observational, cross-sectional study used US Centers for Disease Control and Prevention data from 2004 through 2019 for county estimates of incidence and prevalence of diagnosed diabetes. County rurality was based on 6 levels (large central metro counties [most urban] to noncore counties [most rural]). Weighted least squares regression was used to relate rurality with diabetes incidence rates (IRs; per 1000 adults) and prevalence (percentage) in adults aged 20 years or older after adjusting for county-level sociodemographic factors (eg, food environment, health care professionals, inactivity, obesity).

Results

Overall, in 3148 counties and county equivalents, the crude IR and prevalence of diabetes were highest in noncore counties. In age and sex ratio-adjusted models, the IR of diabetes increased monotonically with increasing rurality (P<.001), whereas prevalence had a weak, nonmonotonic but statistically significant increase (P=.002). Further adjustment for sociodemographic factors including food environment, health care professionals, inactivity, and obesity attenuated differences in incidence across rurality levels, and reversed the pattern for prevalence (prevalence ratios [vs large central metro] ranged from 0.98 [95% CI, 0.97 to 0.99] for large fringe metro to 0.94 [95% CI, 0.93 to 0.96] for noncore). In region-stratified analyses adjusted for sociodemographic factors including inactivity and obesity, increasing rurality was inversely associated with incidence in the Midwest and West only and inversely associated with prevalence in all regions.

Conclusion

The crude incidence and prevalence of diagnosed diabetes increased with increasing county rurality. After accounting for sociodemographic factors including food environment, health care professionals, inactivity, and obesity, county rurality showed no association with incidence and an inverse association with prevalence. Therefore, interventions targeting modifiable sociodemographic factors may reduce diabetes disparities by region and rurality.