Hospital competition, quality, and cost of prostate cancer care.

235 Background: Hospital competition is important for addressing the disparity in quality and cost of prostate cancer care. Study objective was to examine the association of hospital competition with process of care (time to treatment, treatment and overuse) and outcomes (medial care use, complications, mortality and cost) in Medicare fee-for-service beneficiaries with prostate cancer. Methods: This was a population-based cohort study of Surveillance, Epidemiological, and End Results-Medicare (SEER-Medicare) data from 1995- 2016, linked with American Medical Association for physician data and American Hospital Association for hospital level data. Eligible patients were men 66 years or older with localized or advanced stage prostate cancer at diagnosis. The Hirschman-Herfindahl index (HHI) was computed for all serving hospitals based on number of competitors, i.e., number of hospitals situated within the hospital referral region(HRR). The Overuse Index (OI) was used to composite measure of overuse during treatment (one year after diagnosis) and follow-up care phase. Outcomes were overall and prostate cancer-specific survival, complications, readmissions, ER visits, and cost. We used survival analysis, including competing risk analysis, Poisson (zero inflated) models for count data, and GLM (log-link) models for cost data. Propensity score and instrumental variable approaches were used to minimize potential biases. Results: In our study cohort of 434,264, 85% of patients had localized disease stage, and 15% had advanced stage. For both localized and advanced stage groups, age, race and ethnicity, geographic region, comorbidity, socio-economic status, and primary treatment differed by hospital competition (high competition vs. low competition). Hospitals within high competition area were more likely to perform surgery, whereas hospitals within low competition area were more likely to perform radiation therapy. Among localized disease patients, low hospital competition was associated with higher hazard of overall mortality (HR = 1.08, 95% CI = 1.07 - 1.10) and prostate cancer-specific mortality (HR = 1.13, 95% CI = 1.09 - 1.17) and higher odds of ER visits (OR = 1.13, 95% CI = 1.11 - 1.15). For advanced stage patients, low hospital competition was associated with higher hazard of overall mortality (HR = 1.11, 95% CI = 1.08 - 1.15) and prostate cancer-specific death (HR = 1.15, 95% CI = 1.09 - 1.18) and higher odds of ER visits (OR = 1.16, 95% CI = 1.11 - 1.22). Higher scores of the OI were associated with higher total medical costs per capita per year, and not associated with overall mortality. Conclusions: This novel study showed that higher hospital competition is associated with improved quality of care (reduced mortality, complications and ER visits) and increased/lower direct medical care cost among patients with localized or advanced stage prostate cancer. Policy measures should be implemented to improve hospital competition.