OBJECTIVES: The Institute for Clinical and Economic Review (ICER) evaluated targeted immunomodulators (TIMs) for moderate-to-severe plaque psoriasis in 2018. Since then, new entrants (i.e.,bimekizumab and deucravacitinib) and new evidence on Psoriasis Area and Severity Index (PASI) outcomes have emerged. The objective was to update ICER’s 2018 systematic review and network meta-analysis (NMA) and compare results from binominal and multinominal NMA approaches.

METHODS: We systematically identified and reviewed randomized controlled trials (RCTs) on TIMs approved for moderate-to-severe plaque psoriasis that reported data on PASI outcomes at the end of an induction period (10-16 weeks). Random-effects Bayesian NMAs were conducted, and results were presented as relative risks (RR). For the binomial approach, we used a generalized linear model with a log link. The model inputs for this approach were the number of patients that achieved a PASI outcome (e.g., PASI 90) and the total number of patients. For the multinomial approach, we created mutually exclusive groups of the PASI data as <50, 50-74, 75-89, and 90-100 and used a multinomial likelihood model with a probit link.

RESULTS: The updated NMA included 73 studies (25 new studies since the 2018 report) of 14 TIMs. The relative rankings of interventions from the binomial and multinomial baseline risk-adjusted models were similar, although RRs for PASI 90 were lower in the binomial model. All TIMs were more efficacious than placebo. However, bimekizumab, the latest entrant, was the most efficacious (Multinomial:81.62 [95% CrI: 68.17-98.62]; Binomial:45.37 [95% CrI: 17.93-51.84]) while apremilast was the least efficacious (Multinomial:14.49 [95% CrI: 10.75-19.42]; Binomial:7.38 [95% CrI: 5.35-8.88]).

CONCLUSIONS: Although perspectives differ on whether the binominal or multinominal approach may provide the best estimate of treatment efficacy, for this analysis, we found minimal differences between the two approaches regarding the relative ranking of the interventions compared to placebo for the PASI 90 outcome.