Matching-adjusted Indirect Comparisons vs Propensity Score Matching with Individual Patient-level Data to Estimate Treatment Efficacy

In lieu of head-to-head trials, indirect treatment comparisons can be performed to compare effectiveness of therapies. A variety of techniques exist including the Bucher method, matching-adjusted indirect comparison (MAIC) with or without Signorovitch weighting, network meta-analyses, and propensity score matching.1 The MAIC method is a technique to compare treatments with adjustment for baseline differences between trials and has been increasing in popularity among comparative effectiveness trials in inflammatory bowel disease (IBD).2,3 Briefly, MAIC assigns a weight to patients from trials with individual patient data (IPD) based on the mean baseline characteristics in trials reporting aggregate data.4 This approach uses propensity scores weighted by the inverse odds of patients with IPD being assigned to the trial with IPD vs the trial with aggregate data.4 However, the presence of heterogeneity between trials in patient characteristics and study design could undermine the validity of indirect comparisons using techniques such as network meta-analyses. We previously published on our comparisons of vedolizumab and adalimumab using propensity score matching of IPD from phase 3 clinical trials programs, which suggested similar magnitude in absolute difference of efficacy when these therapies were compared, as was found in the randomized controlled trial, VARSITY.5 The purpose of this report was to determine whether MAIC methodology would have also arrived at similar conclusions.