OBJECTIVES: Traditional oncology Partitioned Survival Models (PSMs) use "treat to progression" approach. Emerging evidence in some trials suggests patients may benefit from treatment beyond progression (TBP). Accurately modelling TBP treatment costs poses challenges to economic evaluations, health technology assessment submissions and reimbursement decisions. We examined manufacturers’ approaches to modelling TBP and evaluated acceptance by National Institute for Health and Care Excellence (NICE).

METHODS: A targeted review of modelling approaches and External Assessment Group (EAG) critique in NICE PSM-based oncology submissions (April 2023 to April 2024) where trial time to treatment discontinuation (TTD) or time on treatment (ToT) exceeded progression-free survival (PFS).

RESULTS: Nine submissions were included. Eight reported clinical benefits (symptom control, slowed progression, absence of later therapy lines, manageable toxicity) as reasons for TBP. Two noted non-clinical reasons (scheduling delays, decision-making processes). Assumptions regarding TBP were employed in every economic model. TTD/ToT was capped to PFS in four submissions, assumed equal to PFS in two, capped by a specified time point in one, and two allowed to exceed PFS. All EAGs critiqued TBP assumptions, considering it a key cost-effectiveness driver and noting that it was important to match costs to the clinical benefits of extended treatment already incorporated into models. In the seven submissions artificially limiting TTD/ToT, EAGs requested scenarios without capping, resulting in significant ICER increases. Contrastingly, in two submissions where TBP was not restricted, EAGs recommended introducing a cap to align with clinical guidelines and ensure consistency. Two submissions were rejected with TBP assumptions contributing to the decisions.

CONCLUSIONS: Correctly incorporating both costs and efficacy impacts of TBP is key for economic evaluation, however, there is no single accepted approach. Guidance from NICE would be beneficial. In its absence, manufacturers should consider the clinical context, seek clinical advice on expected real-world use, consider precedent, and explore scenarios with and without capping.