Presentation (CTI)

OBJECTIVES: Population-adjusted indirect comparisons (PAICs) are increasingly used to support health technology assessment (HTA) submissions in the absence of head‑to‑head evidence. This study includes an update of previous work and aims to synthesize recent methodological developments in PAICs and evaluate their uptake, performance, and acceptability within the National Institute for Health and Care Excellence (NICE) appraisals.
METHODS: A targeted literature review was conducted, covering methodological studies published from 2022 to May 2026. Studies that proposed, extended, or evaluated PAIC methods were included. In addition, a review of NICE oncology appraisals since 2022 was performed to examine the PAIC methods used and the associated critiques from the Evidence Assessment Groups (EAGs).
RESULTS: A total of 25 studies were included, of which nine (36%) proposed new PAIC approaches [e.g., doubly robust estimators, regression‑augmented matched-adjusted indirect comparison (MAIC), and network meta‑interpolation]. Four studies (16%) proposed methodological extensions [e.g., quantitative bias analysis and improved effective sample size (ESS) diagnostics]. Ten studies (40%) included formal method comparisons, consistently showing that traditional MAIC is unstable under poor covariate overlap, while hybrid and doubly robust estimators improve bias and variance performance but remain sensitive to unmeasured confounding. Across 27 NICE appraisals, MAIC was the most frequent method used (20/27; 74%). PAICs were often considered acceptable given the available data; however, EAGs frequently judged the resulting evidence as highly uncertain. Key concerns included residual confounding due to unmeasured or unobserved effect modifiers, reduction in ESS, poor covariate overlap, and strong unverifiable assumptions in unanchored designs.
CONCLUSIONS: Despite rapid methodological advances, PAIC methods remain subject to scrutiny in HTA decision‑making. Future progress requires standardized frameworks that justify the choice of the baseline adjustment, integrate robust estimators (e.g., doubly robust approaches) with formal quantitative bias analysis, improved overlap diagnostics, and transparent reporting to better characterize uncertainty and improve credibility in HTA evaluations.