Objective: To test alternative designs of outcomes-based contracts (OBCs) in terms of their impact on two key uncertainties: long-term value and short-term budget impact.

Background: CAR-T therapy is a promising cellular therapy that presents financial challenges for all players of the health system. It is available for patients with childhood B-cell acute lymphoblastic leukemia or aggressive B-cell non-Hodgkin’s lymphoma for whom other treatment options have failed. The importance of studying different payment models grows as gene and cell therapies expand beyond rare cancers to more common tumors.

Methods: The Institute for Clinical and Economic Review’s report shows Tisagenlecleucel and Axicabtagene ciloleucel to be cost-effective compared to the standard chemotherapy while being subject to enormous uncertainty over long-term survival outcomes. Using our replicated CAR-T model, we test three payment vehicles: (1) a responder-based model (RBM); (2) a simple-discount model (SDM); and, (3) an annuity-based model (ABM) in terms of their impact on the short-run budgetary uncertainty and long-term value of treatment.

Results: Our analysis demonstrates that: the RBM reduces uncertainty in value related to the lack of generalizability of the trial results to the real world, but that uncertainty in value related to long-term survival remains; the SDM gives the least budgetary uncertainty, but the greatest uncertainty in value since no adjustment is made for observed outcomes; the ABM gives the greatest reduction in uncertainty related to value at the expense of increasing budgetary uncertainty.

Conclusion: Few studies have compared the effects of alternative OBC designs on stakeholder uncertainty and to date most OBCs have focused on responder type designs. We show that the responder design, minimises neither budgetary or value uncertainty compared to alternative designs, but that a reduction in budgetary uncertainty and value uncertainty cannot be achieved simultaneously.