Geographic variation in the incidence of treated end-stage renal disease

To facilitate identification of geographic clusters of areas with high or low incidence of treated end-stage renal disease, the 1983 to 1988 incidence by county was studied among whites and nonwhites less than 60 yr of age in the United States. End-stage renal disease incidence counts for 1983 to 1988 were obtained from the United States Renal Data System data base and linked to the 1985 county population obtained from U.S. Census data. Maps were smoothed by the method adopted by the National Cancer Institute that smooths only according to variability of the local rates, ignoring geographic information on clustering of events. In addition to identifying specific counties with exceptionally high or low incidence, geographic patterns were observed with many similarities across whites and nonwhites: notably high rates of disease in areas of the Southwest, the Southeast and in counties with Native American reservations and low rates in the West and Northwest. On the basis of these findings, several hypotheses are presented to explain the observed variation in treated end-stage renal disease incidence rates.