PHARMACOECONOMIC MODELING OF BIOSIMILARS IN THE US: A CONCEPTUAL FRAMEWORK

Purpose: The impact of introducing biosimilars to the US market is uncertain due to pricing, regulatory actions, and market penetration. Developing budget impact (BI) and cost effectiveness (CE) models for these products will aid decision-makers, but will pose unique challenges. We developed a conceptual framework to provide guidance in modeling biosimilars in the US setting.

Methods: We identified key challenges in modeling biosimilars based on a targeted literature search and review. We leveraged existing modeling methodology, experiences from the US generics and EU biosimilars markets, and expert opinion to establish recommendations for addressing these challenges.

Results: Three main sources of uncertainty, and recommendations for overcoming each, were established (Figure):

1.     Price;

2.     FDA decisions regarding interchangeability and indications;

3.     Market share.

With uncertainties about future biosimilar pricing, model price estimates could be based on product- and setting-specific predictive modeling; alternatively, models could assume an average 35% biosimilar discount 10 years post-market entry. Models will depend on FDA decisions regarding biosimilar interchangeability and indications. If a biosimilar is not deemed interchangeable, modelers need to assume reduced future market share, add administrative costs of pharmacist substitution requests (e.g., prior authorization; prescribers notifications), and address adverse event data reliability if pharmacovigilance is a concern. If biosimilar indications differ from those of the reference biologic, modeling off-label biosimilar use may not be appropriate. With biosimilar market penetration uncertain, market share estimates should include the impact of price, interchangeability, and indications. In the absence of available data, an average 60% biosimilar penetration achieved 10 years post-market entry could be assumed. Explicit sensitivity analyses with wide confidence intervals for price and market share should be conducted in all biosimilar models.

Conclusions: While uncertainties exist in developing BI and CE models for any new product, the interacting components involved in biosimilars make modeling particularly challenging. This framework provides concrete recommendations for addressing these challenges in modeling biosimilars and will serve to ultimately aid US payer, provider, and policy decision-making.