Abstract:
|
Randomized clinical trials (RCTs) are the gold standard for estimating treatment effects, but the trial findings may not be applicable to target populations of interest. Propensity Score (PS)-based methods have been shown to mitigate this issue. For example, inverse probability of selection weighting (IPSW) can be used to reweight the trial sample as a function of the PS, allowing the trial sample to resemble the target population. Missing data in covariates used in the PS estimation can threaten the validity of such methods, however. Limited work has been conducted to address missing data in this context. Multiple Imputation (MI) is a well-established and accessible method for handling missing data. Unfortunately, there is no consensus on best statistical practice for utilizing MI in estimating and integrating the PS when generalizing RCTs. We conducted an extensive simulation study to evaluate properties of estimators under a variety of MI strategies that fall under two umbrellas (passive and active), coupled with two general strategies for integrating PS into analyses (within and across). We illustrate considerable heterogeneity across methods and provide practical guidelines.
|