Improving RCT-Based CATE Estimation Under Covariate Mismatch via Double Calibration

TL;DR

This paper introduces MR-OSCAR, a novel two-stage estimator that enhances heterogeneous treatment effect estimation by combining observational and randomized trial data, even with covariate mismatch, through imputation and calibration.

Contribution

The paper proposes MR-OSCAR, a new method that improves causal effect estimation under covariate mismatch by integrating imputation and calibration, with theoretical guarantees and practical validation.

Findings

Imputation improves estimation when the observational data predict missing variables well.

Borrowing information is most beneficial with strong predictability and moderate trial size.

The method outperforms naive approaches in simulations and real-world applications.

Abstract

We develop estimators that improve precision of heterogeneous treatment effect estimates that allow borrowing information from observational studies when the available covariates in each data source do not perfectly match. Standard data-borrowing methods often assume perfectly matched covariates. We propose MR-OSCAR, an RCT-calibrated, two-stage estimation approach that first predicts the trial-missing variables using the observational data via imputation and then calibrates observational outcome predictions to the randomized trial, preserving the causal contrast, unlike the results for generalization, where imputation does not improve performance. Our theory gives finite-sample guarantees with a transparent error decomposition including an imputation error that shrinks as the observational mapping becomes more predictable. Simulations show that imputation almost always outperforms naively using only the shared covariates and clarifies when borrowing helps (strong predictability of the missing block, moderate trial size) and when it does not (poor predictability or dominant trial-only moderators). We motivate the approach with the Greenlight Plus trial on early childhood obesity and outline a forthcoming EHR analysis at Vanderbilt, highlighting the use of our method in common scenarios where data do not perfectly align.