Projected impact of oncology biosimilar substitution from the perspective of provider risk in value-based oncology payment models.

Abstract

e18836 Background: Oncology biosimilars may play an important role in managing risk for providers participating in value-based payment (VBP) models. The impact of biosimilar substitution on risk to providers remains unclear, as prior researchers have adopted a generic budget impact approach. Methods: We estimated the impact of biosimilar substitution on financial risk to providers in oncology VBPs by applying simulation approaches to model the quantitative methodology (e.g. episode framing, pricing, risk adjustment) of Medicare’s Oncology Care Model (OCM). Patient demographic and utilization data to fit the models were drawn from the Medicare Limited Data Set (LDS). Risk was defined as total cost of care (TCOC) relative to target price for an episode per the OCM methodology. Target prices were estimated using the most recently available OCM risk adjustment coefficients (2020). Biosimilars for six oncology agents were examined: bevacizumab, rituximab, trastuzumab, epoetin alfa, filgrastim, and pegfilgrastim (hereafter biosimilar investigation agents – BIAs). Episode TCOC was computed under the assumption of the use of reference BIAs and was then recomputed after biosimilar substitution. The study population consisted of 1620 episodes framed per the OCM methodology using the most recently available LDS data (2019-2020) that had use of a BIA. The impact to provider groups was estimated by computing the change in OCM risk bands (incentive payment earned/ neutral/ payment owed back to Medicare) associated with biosimilar substitution in panels of 100 randomly selected episodes from the study population over 10000 simulation runs. In our base model, we assumed that substitution for antineoplastics would occur only in treatment naïve patients and that substitution for supportive therapies would occur for all eligible patients. Treatment naïve status was assessed by examining longitudinal LDS data from 2016-2020 for beneficiaries from the study population. Results: Biosimilar substitution resulted in a mean reduction in cost relative to target of approximately $1,200 per eligible episode and reduced the proportion of practices that were above benchmark for eligible episodes by 33% and increased the number of practices below target for eligible episodes by 42%. Additional scenario analyses suggested that adoption strategy was a major determinant of potential impact. If assumptions of antineoplastic substitution requiring treatment naïve status were relaxed, costs savings relative to target would go from $1200 per eligible episode to $2700. Conclusions: Biosimilar substitution significantly reduces aggregate provider risk in OCM, representing a significant potential intervention for providers to mitigate risk in oncology VBPs both in terms of absolute costs saved relative to target and reduction in risk band relative to payer projections.