Multiple pathological processes are responsible for renal transplant loss. Development of effective prophylactic regimens requires that risks associated with different outcomes be assessed when several processes act together. A multistate competing risk model that derived multiple endpoint probabilities without averaging impacts of covariates over different outcomes or overestimating fractions of survivors was used to estimate the influence of 16 factors simultaneously on four failure types (acute rejection, chronic rejection, death with a functioning graft and all other causes combined). We analyzed 12,413 non-Black and 5,121 Black patients undergoing primary cadaveric kidney transplantation reported to UNOS during 1994-96. Black recipients generally demonstrated higher rates of failure, but the method provided joint probabilities for each failure type and covariate level. To illustrate, a middle-aged Black recipient had probabilities of 6.5%, 5.9%, 2.6% and 6.6% to experience acute rejection failure, chronic rejection failure, death with a functioning graft and other failures, respectively, within 3 years. The corresponding non-Black's chances were 2.9%, 2.5%, 2.7% and 3.6%. The overall 3-year survival equaled 78.4% [100%-(6.5% + 5.9% + 2.6% + 6.6%)] and 88.3% for Blacks and non-Blacks, respectively. The growing complexity of transplantation will benefit from new analytic approaches that accommodate multiple transitions along diverse paths to eventual graft loss or death.