Association between health-care system and prostate cancer mortality for African American men with localized and metastatic prostate cancer.

Abstract

27 Background: African American (AA) men with prostate cancer (PC) present with more advanced disease and have worse survival than comparable non-Hispanic White (White) men. Recent studies suggest that receiving care within an equal access setting may attenuate these disparities. We hypothesize that AA men receiving care within the Veterans Health Administration (VHA) will have improved outcomes compared to AA men receiving care in the general population as assessed by the Surveillance, Epidemiology, and End Results (SEER) database. Methods: We identified AA men diagnosed with PC between 2004 and 2015 in the VHA and SEER. For comparisons of covariate distributions across subgroups, we used the chi-squared test with continuity correction. We analyzed the cumulative incidence (with 95% confidence intervals (CIs)) of PC specific mortality (PCSM) in the VHA and SEER. Additionally, multivariable Cox proportional hazards models controlling for demographic information were performed. Results: The cohort included 85,409 AA men (VHA: 27,415, SEER: 57,994). Median follow-up was 4.79 years in the VHA and 5.16 years in SEER. In the VHA, AA men were more likely to present with localized disease (VHA 94.7% vs SEER 86.4%, p < 0.001) and less likely to have metastatic disease (3.2% vs 4.3%, p < 0.001). The 5-year cumulative incidence of PCSM was lower for patients in the VHA (VHA: 3.8% [CI: 3.5-4.1%] vs. SEER: 5.0% [CI: 4.8-5.2%], p < 0.001). The PCSM difference was largest in men with metastatic disease. In metastatic patients, cumulative incidence of PCSM at five years was significantly lower in the VHA (VHA 52.5% [CI: 48.0-56.5%] vs. SEER 64.8% [CI: 62.3-67.1%], p < 0.001). In contrast, AA men with localized disease had similar PCSM in the VHA and SEER (VHA 2.4% [CI: 2.2-2.6%] vs. SEER 2.6% [CI: 2.4-2.7%], p = 0.09 at five years). On multivariable analysis, VHA system was associated with lower PCSM [Hazard Ratio (HR): 0.91, p < 0.001]. There was a significant interaction between VHA system and distant metastases at diagnosis [p < 0.001] indicating larger differences in PCSM by healthcare system in metastatic patients as compared to localized patients. VHA system was associated with reduced PCSM in metastatic patients [HR 0.84, p < 0.001] but not in localized patients [HR 0.96, p = 0.13]. Conclusions: AA men in the VHA had a significantly lower incidence of PCSM than those in the SEER database, especially for those who presented with distant metastases at diagnosis. Future work should examine how cost and access to care affect disparities in outcomes for AA men.