OBJECTIVES : In the absence of Head-to-head clinical trials population adjusted ITC such as matching adjusted indirect comparison (MAIC) and STC can account for between-trial imbalances in the distribution of effect modifiers. STC is a regression-based approach which predicts the outcome in the target population. As ITC conducted in a Bayesian setting has been increasing in popularity, our objective is to bring together the flexibility of Bayesian modeling with STC model by NICE and compare the output.

METHODS : We simulated the hypothetical data from three trials (treatment B vs. A) and a comparator trial (treatment C vs. A), using R package Wakefield RESULTS : Treatment effect estimates obtained from the Bayesian model after adjusting for the different baseline risk were logOR B vs. A: -2.49 (-3.06, -1.95). These results were aligned with the pooled estimates obtained from STC (as per the NICE methods) logOR B vs. A: -2.45(-2.99, -1.90).

CONCLUSIONS : Bayesian IPD regression with study level baseline terms added in the outcome model can be used to combine IPD from multiple trials when population adjusted ITCs are required. Our findings showed that STC results from both frequentist and Bayesian frameworks are aligned with the latter offering more flexibility. STC can also be an option to explore when MAIC is not feasible due to small effective sample size.