A Framework for Reliable, Transparent, and Reproducible Population-Adjusted Indirect Comparisons

Population-adjusted indirect comparison (PAIC) methods aim to address some of the potential shortcomings of conventional approaches to indirect treatment comparisons by adjusting for imbalances in effect modifiers or prognostic factors and allowing for unanchored indirect treatment comparisons from disconnected networks of evidence. Health technology assessment bodies have published guidance and best practice recommendations for PAICs. However, recently published reviews of published PAICs have highlighted notable variability in implementation and a lack of transparency in the decision-making process in analyses and reporting; this hinders the interpretation and reproducibility of analyses, which, in turn, could affect reimbursement decision-making. We propose a systematic framework to address these challenges by describing considerations on six key elements of analyses: (1) definition of the comparison of interest (e.g., in terms of an estimand), (2) selection of the PAIC method, (3) selection of adjustment variables, (4) application of adjustment method, (5) risk-of-bias assessment, and (6) comprehensive reporting.