Patient and population impacts of multigene panel and pembrolizumab coverage in metastatic melanoma

Background: Targeted treatment or immunotherapy may yield increased, durable responses for melanoma patients. Whether patient-level benefits translate to population health is unknown. This study sought to estimate patient and population impacts of a cancer control policy that reimbursed multigene panel testing and pembrolizumab for metastatic melanoma in British Columbia, Canada. Methods: This retrospective study examined a population-based cohort of 721 adults diagnosed with metastatic melanoma in British Columbia who received single or multigene testing between 2013 and 2018. We determined patient-level policy impacts using 1:1 genetic algorithm matching of policy-affected patients with historical control patients and Kaplan–Meier analysis and inverse probability of censoring weighted regression of 2-year health-care costs and survival times. For population-level effects, we applied interrupted time-series analysis on monthly health-care system expenditures and mortality rates, estimating autoregressive integrated moving average and generalized least squares Poisson regressions. Results: Matched cohort analysis (control patients, n ¼ 154; intervention patients, n ¼ 154) found mean cumulative patient-level cost increases of CAD$53 963 (95% confidence interval [CI] ¼ $35 641 to $72 621; P < .001) and increased survival times of 111 days (95% CI ¼ 44 to 166 days; P < .001) over 2 years. Higher patient-level systemic therapy spending of CAD$48 890 (95% CI ¼ $31 110 to $66 910; P < .001) drove overall cost differences. Population-interrupted time-series analysis detected an immediate, sustained increase in mean monthly health-care expenditures of CAD$1921 (95% CI ¼ $935 to $2908; P < .001) per patient. Higher overall spending did not coincide with population-level mortality changes. Conclusions: The policy of reimbursing multigene testing and pembrolizumab produced patient survival improvements, but selectivity of response prevented population mortality improvement. Health-care system costs statistically significantly increased at the patient and population levels.