Skeletal-related events and mortality among men diagnosed with advanced prostate cancer: The impact of alternative measures of radiation to the bone

PURPOSE/OBJECTIVE(S): Skeletal-related events (SREs), which include radiation to the bone (RtB), can occur among patients with bone metastasis (BM). There is a recognized potential for misclassification of RtB when using claims data. We compared alternative measures of RtB to better understand their impact on SRE prevalence and SRE-related mortality. METHODS AND MATERIALS: We analyzed data for stage IV prostate cancer (PCa) cases identified between 2005 and 2009 in the Surveillance, Epidemiology, and End Results registry linked with Medicare claims. We created two measures of RtB: 1) a literature-based measure requiring the presence of a prior claim with a BM code; 2) a new measure requiring either that the BM code coincided with the radiation episode or that the duration of the radiation episode was less than or equal to 4 weeks. We estimated adjusted hazard ratios of an SRE using both measures among stratified samples: no metastasis (M0), metastasis to bone (M1b) and other sites (M1c). RESULTS: The study sample included 5,074 men with stage IV PCa (median age 77 years), of whom 22% had M0, 54% had M1b, and 24% had M1c disease at time of PCa diagnosis. Based on Approaches 1 and 2, the proportion with probable RtB was 5% and 8% among M0, 30% and 30% among M1b, and 25% and 27% among M1c patients. Among M0 patients, the adjusted hazard ratio (AHR) associated with an SRE was 1.27 when using Approach 1 (95% confidence interval, CI: 0.95–1.7) and 1.49 when using Approach 2 (95% CI: 1.14–1.96). However, the impact of SREs on mortality did not differ between both approaches among M1b and M1c patients. CONCLUSION: We found that alternative measures used to define RtB as SRE in claims data impact conclusions regarding the effect of SREs on mortality among M0 but not M1 patients.