Structured expert elicitation of long-term survival in non-small cell lung cancer: a case study applying guidance from NICE Technical Support Document 26

Background

Extrapolation of overall survival in health technology assessment (HTA) is a key area of uncertainty, as equally well-fitting models can generate divergent lifetime projections. This can lead to significant implications for decision-making. Subject experts are often consulted on the plausibility of survival projections, but input is rarely obtained in a robust manner.

Objective

This case study applies guidance from the National Institute for Health and Care Excellence Decision Support Unit Technical Support Document (TSD) on using structured expert elicitation for long-term survival estimation (TSD 26). This study focuses on its feasibility of implementation from the perspective of a company preparing a HTA submission.

Methods The REVEL trial, evaluating ramucirumab plus docetaxel for stage IV non-small cell lung cancer, was selected as a case study. The elicitation was based on the Sheffield Elicitation Framework (SHELF) and undertaken in accordance with TSD 26.

Results

Five experts from four countries attended across three virtual workshops. Experts provided quantitative judgements on overall survival at 4 years for the intervention and comparator from the REVEL trial. Qualitative discussion of the change in hazard and scenario testing demonstrated the internal consistency of experts’ judgments and provided valuable context for expert judgements. Group discussion in separate workshops posed challenges when the facilitation team had to relay the views of absent experts.

Conclusions

This case study demonstrates that it is feasible to follow the recommendations provided in TSD 26 when conducting elicitation for long-term survival outcomes and highlights the critical role of discussing the hazard function in group discussions.