A spatial mixed model with a Poisson distribution, county-specific random intercepts, and adjustment for age, sex, race-ethnicity, and Medicare Part B coverage was fit to calculate the risk-standardized rate of ischemic stroke hospitalization for each county. Smoothed rates were obtained for both risk-standardized stroke hospitalization (A) and the percentage of Medicare beneficiaries who were dual eligible (B). Counties or county-equivalents were grouped into 25 quantiles based on these data and shaded according to a gradient from green (lowest percentage dual eligible/lowest hospitalization rate per 100,000 beneficiary-years) to red (highest percentage dual eligible/highest hospitalization rate per 100,000 beneficiary-years). BY, beneficiary-years.