In the United States, the age-and-gender-adjusted incident rate of end-stage renal disease (ESRD) for blacks has been 4 times higher than that for whites. We analyzed patient information and medical services contained in the Medicare 5% random sample database. White (n = 977,436) and black (n = 77,800) Medicare enrollees who were at least 65 years old on January 1, 1997, were followed from 1999 to 2001. Hierarchical Cox regression models were used to estimate the relative risk of ESRD for blacks (with reference to whites) after adjustment for age and gender, socioeconomic status, special health conditions (anemia, chronic obstructive pulmonary disease, cardiovascular disease), primary causal diseases of ESRD (eg, diabetes, hypertension), diabetes care and preventive care (eg, hemoglobin A1c or lipid testing), and physician visits for primary or specialty care. The relative risk of ESRD for blacks (with reference to whites) was 3.52 (95% confidence interval [CI], 3.25-3.80) after adjustment for age and gender; 2.90 (95% CI, 2.67-3.15) after adjustment for socioeconomic status and special health conditions; and 2.11 (95% CI, 1.94-2.30) after further adjustment for primary causal diseases of ESRD, diabetes care and preventive care, and physician visits. We conclude that a higher prevalence of primary causal diseases of ESRD and lower access to diabetes care, preventive care, and primary physician visits in blacks compared with whites partially accounts for the racial difference in the incidence of ESRD in the elderly Medicare population. Public health policy should focus on improving access to care, which may lower the burden of ESRD in minority and other at-risk populations.