OBJECTIVES: External evidence is often used in health technology assessment (HTA) submissions to NICE to supplement data available from the pivotal clinical trial programme when performing survival extrapolation; however, there is no published guidance (including from NICE) regarding how this should be done. This study aimed to investigate how external evidence has been used in recent appraisals of cancer drugs conducted by NICE to aid extrapolation of survival outcomes.

METHODS: The NICE website was searched to identify non-terminated appraisals of cancer drugs published in 2021. Committee papers were searched for relevant information covering the: survival modelling approach taken by the company for the intervention and its comparator(s); independent external assessment group critique; and related commentary from the committee.

RESULTS: In total, 36 appraisals were identified. External evidence sources used included national statistics, registry data, observational studies, and data from other clinical trials. Most appraisals (n=30) included at least one source of external evidence as part of the estimation of survival for either the intervention or a comparator, and the majority of appraisals (at least n=23) factored background mortality estimates into extrapolations. Excluding background mortality adjustments and indirect treatment comparisons as sources of external evidence, around one-third of appraisals (n=13) integrated some other external evidence source into survival modelling, including data from another clinical trial (n=6), non-trial (e.g., observational study) data (n=6), and clinical expert opinion (n=3).

CONCLUSIONS: Several types of external evidence have been used to inform recent NICE appraisals. However, few appraisals factored external evidence from either a clinical trial or a non-trial source into the survival model fitting process. Technical descriptions of how external evidence was incorporated were often limited, including for background mortality adjustment. Further research is required to understand how, when, and why external evidence should, or should not, be used.