Presentation (CTI)

OBJECTIVES: Indirect treatment comparisons (ITCs) are increasingly used when head-to-head evidence is unavailable. A feasibility assessment (FA) evaluates whether the available studies, data, and analytical assumptions can support a valid comparison. Inadequate FA reporting may prevent researchers and decision-makers from evaluating the credibility of resulting estimates. This study assessed how frequently and comprehensively FA methods and key feasibility components were reported in published ITCs.
METHODS: PubMed was searched from inception to June 2026 for peer-reviewed ITCs, including network meta-analyses (NMAs), matching-adjusted indirect comparisons (MAICs), and multilevel network meta-regressions (ML-NMRs). Open-access full texts were reviewed to determine whether ITC feasibility was explicitly assessed and whether 17 predefined feasibility-reporting components were reported. These comprised 14 general components, including clinical and methodological comparability, effect modifiers (EMs), prognostic factors (PFs), transitivity, network connectivity, and data availability, and 3 additional population-adjusted ITC-specific components: IPD-aggregate data compatibility, covariate availability, and population overlap. Reporting frequencies were summarized overall and by ITC method.
RESULTS: Of 807 publications, 567 open-access full texts were available, including 368 NMAs, 195 MAICs, and 4 ML-NMRs. Of these, 349 (61.6%) mentioned conducting an FA, among which methodological details were often brief or embedded within the description of the analytical approach rather than presented as a distinct FA. Evidence-base descriptions (99.7%), study exclusions (95.4%), connectivity (89.4%), and population, treatment, outcome, and study design comparability (82.5%-88.5%) were commonly reported. Transitivity (56.7%), EM distributions (48.7%), common comparator suitability (48.4%), and PFs (43.3%) were less frequent. Among the 14 general FA components, mean reporting completeness was 10.0 for NMAs (n=300), 12.6 for MAICs (n=46), and 12.3 for ML-NMRs (n=3).
CONCLUSIONS: Inconsistent and incomplete FA reporting limits assessment of whether published ITCs are methodologically justified and sufficiently robust for interpretation or downstream use. Standardized reporting would strengthen critical appraisal, support replication of analytical judgments, and increase confidence in relative effect estimates.