Background: Limited information is available regarding the association between race/ethnicity, health service use, cost, and mortality in older patients with advanced (T3 or greater) stage prostate cancer. The objective is to analyze the race/ethnic differences in mortality, cost, and assess the mediating effect of prostate cancer screening and treatment on these differences in fee-for-service Medicare patients with advanced prostate cancer Methods: Retrospective, observational, case-control study using SEER-Medicare linked data. Cohort consisted of 15,054 elderly men diagnosed with advanced stage prostate cancer between 2001and 2004 and followed retrospectively for up to 2009. Cancer-free controls from Medicare data were used to determine the incremental cost of advanced stage prostate cancer. Racial/ethnic variation in health service use, cost, and mortality were analyzed using Poisson, GLM log-link, and Cox regression models. Results: For the age 66 to 75 and 76 to 85, age groups, racial/ethnic differences in health service use, cost, and all cause mortality were observed. Blacks were less likely to have received prostate cancer screening in the year prior to diagnosis of advanced prostate cancer, less likely to have received any treatment after diagnosis and had higher disease specific mortality. After adjusting for prostate cancer screening and treatment across age groups however, odds of prostate cancer-specific mortality were comparable between racial/ethnic groups suggesting that screening may have some mitigating effect on outcomes in the AA population HR 1.23 (1.07-1.41) versus HR 1.12(0.98-1.29). Conclusions: The pattern of racial/ethnic disparity varies by age group with higher mortality among black men. This may be attributable to disparity in prostate cancer screening or treatment. This suggests that the lack of consideration for racial considerations in the U.S. Preventive Services Task Force PSA recommendations could have disparate racial impact.