Hybrid Control Design with Commensurate Prior Constructed Using Propensity Score-Matched External Controls: A Simulation Study

A hybrid control design is a way to borrow information from external controls to augment concurrent controls in randomized controlled trials. A major challenge in the hybrid control design is its inability to eliminate prior-data conflict caused by systematic imbalances in measured and/or unmeasured confounders between patients in the current trial patients and external controls. To prevent the prior-data conflict, a combined use of propensity score matching and Bayesian commensurate prior has been proposed in the context of hybrid control design; however, practical aspects of the method have not been fully addressed. We conducted a comprehensive simulation study to evaluate operating characteristics of the combined use of propensity score matching and Bayesian commensurate prior, and revealed some key features, such as the choice of the propensity score matching procedure, the size of the external control pool, and the choice of the prior of commensurability parameter, that practitioners should be careful in the implementation. We developed an R package “psBayesborrow” that is freely available at CRAN to facilitate the use of the methods addressed. Supplementary materials for this article are available online.