Outcome Regression Methods for Analyzing Hybrid Control Studies: Balancing Bias and Variability

TL;DR

This paper reviews and compares outcome regression methods for hybrid control studies, highlighting their bias-variance trade-offs and demonstrating their effectiveness through simulations and real-world examples.

Contribution

It introduces and evaluates three outcome regression-based methods for hybrid control studies, emphasizing their bias-variance trade-offs and practical performance.

Findings

Weighted regression performs favorably compared to other model-based methods.

Simulation results show different bias-variance trade-offs among the methods.

Real-world examples demonstrate practical applicability.

Abstract

There is growing interest in a hybrid control design in which a randomized controlled trial is augmented with an external control arm from a previous trial or real world data. Existing methods for analyzing hybrid control studies include various downweighting and propensity score methods as well as methods that combine downweighting with propensity score stratification. In this article, we describe and discuss methods that make use of an outcome regression model (possibly in addition to a propensity score model). Specifically, we consider an augmentation method, a G-computation method, and a weighted regression method, and note that the three methods provide different bias-variance trade-offs. The methods are compared with each other and with existing methods in a simulation study. Simulation results indicate that weighted regression compares favorably with other model-based methods that seek to improve efficiency by incorporating external control data. The methods are illustrated using two examples from urology and infectious disease.