OBJECTIVES : Manufacturers of immune checkpoint inhibitors (ICI) are increasingly seeking reimbursement based on immature trial evidence on overall survival (OS), which causes challenges for cost-effectiveness modelling and raises concerns for payers. This study aims to summarise approaches adopted for oncology cost-effectiveness models for ICIs submitted to the National Institute for Health and Care Excellence (NICE) when the modelled OS for the intervention is only partially based on data from the pivotal trial.

METHODS : NICE technology appraisals (TAs) for ICI published between 2014 and 2019 were identified through January 2020. TAs with the modelled OS only partially sourced from the pivotal trial were then further screened in a peer-reviewed process.

RESULTS : Eight TAs across multiple oncology indications were identified; the pivotal trial median follow-up, when reported, was below 20 months. For six of the TAs (319, 357, 366, 384, 400 and 600), the manufacturers modelled OS using pivotal trial data where available and adopted external data for the remainder of the time horizon. Criticisms from the Evidence Review Group (ERG) were raised related to: complexity and lack of transparency of the approaches, heterogeneity between external sources and trial populations, and clinically implausible behaviour of mortality risk. Three TAs (319, 462 and 540) adopted model frameworks which included additional health states to reflect the impact of subsequent therapies on survival. Despite the unavailability of mature OS data, all the identified TAs received positive guidance recommendation, with two within the Cancer Drugs Fund.

CONCLUSIONS : Manufacturers of ICIs appear to use similar methods and modelling approaches and rely on external data sources to inform modelling of long-term survival when trial data are not mature enough to be used as the sole source. Despite these limitations, NICE has given positive guidance recommendation in all TAs identified.