OBJECTIVES: After introducing biosimilars, expectations were high for dramatic cost reductions for biologics. However, savings have been modest in the US. This may be due to the US reimbursement model. For physician-administered drugs, which includes many biologics, providers receive payment based on the percentage of the cost billed, incentivizing the highest-cost product. Currently, US policymakers are debating changing this policy. However, to understand changing the system, causal estimates of the savings seen in the current system are needed, which this study aims to provide.

METHODS: To model cost-savings associated with biosimilar introduction, we used Bayesian structural time series models (BSTS), which utilize information from control time series data that closely resembles biologics’ ASP trends before biosimilars’ market entry. We identified prices from Medicare Part B Drug Average Sales Price (ASP) Pricing datasets (2005 Q1 - 2024 Q2). We included the following biologics: bevacizumab, epoetin(ESRD and non-ESRD), filgrastim, infliximab, peg-filgrastim, rituximab, and trastuzumab. We identified a set of control medications for each biologic such that the controls’ ASP trends closely resemble those of their respective biologics and are unlikely to be influenced by biosimilars’ market entry. We generated the predicted counterfactual ASPs, accompanied by 95% prediction intervals. Since the included biologics had different post-biosimilar entry times, we reported the cumulative price difference between biosimilars and the corresponding parent biologic over 3-year, and all post-biosimilar times.

RESULTS: Except for filgrastim, all biologics had a significant decrease in their ASP per dose values after the entry of biosimilars. Three years after biosimilar entry, most biologics ASP per dose decreased by 10 to 19%, with epoetin (ESRD) and epoetin (non-ESRD) having the greatest reduction in ASP per dose: -19.3% [-25.1%, -6.3%] and -19.6% [-25.5%, -6.7%], respectively.

CONCLUSIONS: Savings from biosimilar introduction are modest. Policymakers should consider alternative reimbursement approaches that align provider and payer incentives.