OBJECTIVES: Population-adjusted indirect comparison methods such as matching-adjusted indirect comparisons (MAIC) and simulated treatment comparisons (STC) are useful tools to correct trial population differences when estimating treatment effects from single-arm trials. However, unanchored MAIC and STC assume that all effect modifiers and prognostic factors are accounted for, which is largely considered impossible to meet. The aim is to address the limitation of the current population adjustment methods where certain prognostic factor and/or effect modifiers are not reported in the comparator trial.

METHODS: STC is a form of outcome regression approach where a statistical model is fitted using the trial population where individual patient-level data are available, and the model is used to predict the outcomes that would have been observed in the aggregate population. We developed an extended STC (ESTC) approach which accounts for unobserved confounding. In ESTC, the regression model also includes the important variables which are not reported in the comparator trial. In the prediction step, a fixed value for the mean of the marginal distribution of the unreported variables is assumed. We proposed a sensitivity analysis to quantify the impact of varying the marginal means of the unreported variables. We also performed a simulation study to assess the performance of the ESTC approach for binary outcomes.

RESULTS: The ESTC approach performed well across all scenario analysis including: varying the magnitude of difference in the marginal means of covariates between trials, varying the strength of the treatment effect and varying the strength of the treatment and covariate interaction. The simulation study shows that the ESTC approach provides an unbiased estimate for treatment effect.

CONCLUSIONS: The ESTC approach formally quantifies the bias associated with unobserved confounding and provides a quantitative assessment of the impact of this bias. This approach increases the robustness of the treatment indirect comparison approach for single-arm trials.