The authors conducted a simulation study to examine whether the race of a patient with end-stage renal disease (ESRD) influences the diagnosis of underlying kidney disease made by the nephrologist. The hypothesis was that ESRD may be more readily ascribed to hypertension in blacks than in whites. Nephrologists practicing in Maryland during 1991 were sent written case histories based on the presentation of seven patients with ESRD. For each case history, the patient's race was randomly assigned to be "black" or "white." The nephrologist's diagnosis of underlying renal disease was recorded as "hypertensive" or "other." Analysis of 197 case histories from 58 physicians (81% of those eligible) was performed using logistic regression. The distribution of underlying causes of ESRD in the case histories was similar to national statistics: hypertensive ESRD, 34%; diabetic ESRD, 30%; glomerulonephritis, 11%; other, 16%; unknown, 10%. Case histories that identified the patient's race as black were more likely (odds ratio = 1.97; 95% confidence interval 1.05-3.68) to result in a diagnosis of hypertensive ESRD than case histories in which the patient's race was said to be white, after adjustment for case history. Analyses that accounted for the physicians' individual tendencies to diagnose hypertensive ESRD yielded similar results. These findings suggest that black ESRD patients may be more likely to be labeled as having hypertensive kidney disease than white ESRD patients with similar clinical histories. Using race as a criterion to establish diagnoses of kidney disease may obscure the interpretation of incidence statistics, affect the management of individual patients, and hinder epidemiologic studies of risk factors for kidney failure. Definition of clear diagnostic criteria for the underlying cause of kidney failure is highly desirable.