Using Propensity Score Weighting to Enhance the Operating Characteristics of Power Prior in Leveraging External Data to Augment a Traditional Clinical Study

The method of power prior has long been used as a tool for leveraging external data to augment a traditional clinical study. More recently, it has been found that integrating propensity scoring into its application has the potential for improved operating characteristics. In this paper, we introduce a new propensity score-integrated power prior strategy which uses propensity score weighting and is distinctive from other such proposals in the literature. This strategy replaces the sufficient statistic in the original expression of power prior with a propensity score weighted version of it. A simulation study shows that the operating characteristics of the proposed weighting strategy compare favorably to those of the original power prior method when there is covariate imbalance, like the stratification strategy we first introduced.