OBJECTIVES: In the absence of direct comparisons between alternative healthcare interventions within cost-utility studies, the French National Authority for Health (HAS) recommends performing network meta-analysis (NMA) to estimate their relative effectiveness. NMA is not always feasible whenever a common comparator between the selected alternatives is lacking. Several indirect pairwise comparisons, such as MAIC and propensity score, can then be performed. The aim of our study was to evaluate how these approaches were accepted in HAS opinions about dossiers currently published.

METHODS: A retrospective analysis of the HAS opinions published until December 31, 2022 was carried out.

RESULTS: 79 dossiers were analyzed from 2020 to 2022. To date, the HAS has not accepted the validity of pairwise comparison methods to assess the efficiency of a new treatment. In the opinions that included several indirect comparisons instead of a NMA, beyond the remarks on the heterogeneity of the studies that questioned the possibility of conducting valid indirect comparisons, the impossibility of constructing an efficiency frontier and therefore of concluding on the efficiency of the assessed product was underlined. In evaluations based on pairwise comparisons, two ICERs will be estimated making it difficult to draw any final conclusion. The HAS suggested checking whether one of the analyses could be considered more relevant and able to support the main analysis.

CONCLUSIONS: The use of pairwise comparisons to obtain a multi-option efficiency frontier was not validated so far by the HAS. In these cases, the HAS recommends that only one comparison be retained to respect methodological validity, despite lack of comparators comprehensiveness in the base case analysis.