Covariate adjustment in subgroup analyses of randomized clinical trials: A propensity score approach

TL;DR

This paper introduces a propensity score weighting method for covariate adjustment in subgroup analyses of RCTs, improving precision and power especially in small subgroups, by balancing covariates with interaction terms.

Contribution

It develops a novel propensity score weighting approach that incorporates covariate-subgroup interactions, enhancing subgroup analysis accuracy in randomized trials.

Findings

Weighted estimators have smaller standard errors than unadjusted ones.

Propensity score weighting matches ANCOVA in efficiency, often better in small samples.

Full-interaction propensity models outperform main-effect models.

Abstract

Background: Subgroup analyses are frequently conducted in randomized clinical trials to assess evidence of heterogeneous treatment effect across patient subpopulations. Although randomization balances covariates within subgroups in expectation, chance imbalance may be amplified in small subgroups and harm the precision. Covariate adjustment in overall analysis of RCT is often conducted, via either ANCOVA or propensity score weighting, but for subgroup analysis has been rarely discussed. In this article, we develop propensity score weighting methodology for covariate adjustment to improve the precision and power of subgroup analyses in RCTs. Methods: We extend the propensity score weighting methodology to subgroup analyses by fitting a logistic propensity model with pre-specified covariate-subgroup interactions. We show that, by construction, overlap weighting exactly balances the covariates with interaction terms in each subgroup. Extensive simulations were performed to compare the operating characteristics of unadjusted, different propensity score weighting and the ANCOVA estimator. We apply these methods to the HF-ACTION trial to evaluate the effect of exercise training on 6-minute walk test in pre-specified subgroups. Results: Standard errors of the adjusted estimators are smaller than those of the unadjusted estimator. The propensity score weighting estimator is as efficient as ANCOVA, and is often more efficient when subgroup sample size is small (e.g.<125), and/or when outcome model is misspecified. The weighting estimators with full-interaction propensity model consistently outperform the standard main-effect propensity model. Conclusion: Propensity score weighting is a transparent and objective method to adjust chance imbalance of important covariates in subgroup analyses of RCTs. It is crucial to include the full covariate-subgroup interactions in the propensity score model.