Facility-Level Variation in Use of Locoregional Therapy for Metastatic Prostate Cancer

Introduction: We sought to understand patient- and institution-level factors associated with use of locoregional therapy for newly diagnosed metastatic prostate cancer in the era before the availability of evidence supporting its efficacy. Methods: We queried the National Cancer Database to identify patients diagnosed with metastatic prostate adenocarcinoma (stage M1) between 2004 and 2017. We assessed patient factors associated with definitive local therapy with radiotherapy or radical prostatectomy using multilevel logistic regression accounting for clustering within institutions. We further characterized trends in facility-level use and examined institutional factors associated with utilization. Results: We identified 35,933 patients with M1 prostate cancer at 1,188 facilities. A total of 4,146 patients (11.5%) received local therapy for M1 disease (radiation therapy in 3,378 and radical prostatectomy in 768). Use of local treatment was concentrated among a smaller number of facilities: 50% of all local therapy was delivered at 161 facilities (14% of total). At the patient level, uninsured status (OR 0.62, 95% CI 0.49-0.79, p<0.01) and high comorbidity (Charlson-Deyo score, OR 0.39, 95% CI 0.26-0.6, p<0.01) were associated with lower odds of local therapy. High-utilizing facilities (top quartile) were more commonly community centers (OR 1.76, 95% CI 10.7-2.95, p<0.01) and differed by geographic region (South Atlantic vs West South Central region: OR 0.48, 95% CI 0.25-0.88, p=0.02). Conclusions: In the period before locoregional therapy was supported by clinical practice guidelines, locoregional therapy use varied significantly at the facility level and was driven by a smaller number of high-utilizing facilities. These findings can contextualize expected increase in the use of local therapy for metastatic prostate cancer.