Indirect Treatment Comparisons With Full Individual Patient Data for Joint Clinical Assessment: Lessons Learned and Best Practices From a Review of HTA Appraisals

OBJECTIVES: More than half of all approvals of new molecular entities (NMEs) in oncology in the European Union (EU) are based on non-randomised evidence. The Member State Coordination Group on HTA (HTACG) has published methodological guidance for quantitative evidence synthesis to support the upcoming introduction of Joint Clinical Assessment (JCA), strongly recommending that indirect treatment comparisons (ITC) of non-randomised data be based on full individual patient data (IPD). The objective of this study was to conduct a review of recent HTA appraisals of oncology NMEs supported by ITC with full IPD to obtain insights into best practices and practical considerations for the conduct of ITCs supporting non-randomised trial evidence for JCA.

METHODS: A search for oncological NMEs approved in the EU based on non-randomised evidence between 2020 and 2023 was conducted in IQVIA’s Market Access Insights database. Retrieved HTA reports from 14 EU countries were assessed by two reviewers and information was extracted on ITC methods used, data sources, agency preferences, critiques, and conclusions on ITCs.

RESULTS: Thirty products were retrieved from the search, 23 of which featured HTA appraisals with IPD-based ITCs. The majority of IPD-ITCs with retrospective designs (e.g., individual sites, disease registries, electronic health record databases, pooled historical data) were not accepted, whereas prospective external control arm designs were largely accepted. Acceptability also differed by HTA agency, with confounding / risk of bias and generalizability being key drivers for non-acceptance. Notably, none of the ITCs in this review were accepted by German and French assessors.

CONCLUSIONS: Health technology developers need to balance methodological considerations and careful ITC design with practical aspects for the implementation of IPD-ITCs to maximise likelihood of their evidence package being positively reviewed in JCA for oncology products based on non-randomised evidence.