OBJECTIVES : Recent years have seen manufacturers of immune checkpoint inhibitors (ICIs) seeking reimbursement with very little or no evidence of overall survival (OS) data. Considered the gold standard oncology endpoint and an important driver of cost-effectiveness, the lack of OS data is often a major concern to payers. This research aims to summarise approaches adopted for oncology cost-effectiveness models for ICIs submitted to the National Institute for Health and Care Excellence (NICE) when data informing OS were not sourced from the pivotal trial in the submitted analysis.

METHODS : NICE technology appraisals (TAs) for ICIs published between 2014 and 2019 were identified through January 2020. TAs without OS informed by pivotal trial data were then further screened in a peer-reviewed process.

RESULTS : Three TAs were identified. Two TAs (553 and 558) had no OS data available from the pivotal trial, while in TA517 the manufacturer presented OS data considered too immature to be used within the cost-effectiveness analysis. In TA553, the manufacturer used additional trial and registry data to inform survival of subsequent lines of therapies: the Evidence Review Group (ERG) raised concerns on the clinical validity of the modelled OS. In TA558, treatment effect on OS was predicted based on recurrence-free survival using a surrogacy equation: the presented scenario utilising subsequent therapies’ survival was preferred by the ERG. In TA517, the manufacturer modelled OS efficacy assuming relative improvement of treatment-naive versus treatment-experienced patients using clinical opinion: the ERG preferred the parametric extrapolation of the immature trial data. All three TAs resulted in recommendations for use within the Cancer Drugs Fund for further data collection.

CONCLUSIONS : In the identified TAs, NICE had a preference for modelling long-term survival using external data, or the pivotal trial data when available, instead of complex regression analyses or surrogate assumptions to inform their decision.