Harmonized Estimation of Subgroup-Specific Treatment Effects in Randomized Trials: The Use of External Control Data
share
Subgroup analysis of randomized controlled trials (RCTs) constitutes an important component of the drug development process in precision medicine. In particular, subgroup analysis of early phase trials often influences the design and eligibility criteria of subsequent confirmatory trials and ultimately impacts which subpopulations will receive the treatment after regulatory approval. However, subgroup analysis is typically complicated by small sample sizes, which lead to substantial uncertainty about the subgroup-specific treatment effects. In this work we explore the use of external control (EC) data to augment an RCT's subgroup analysis. We define and discuss harmonized estimators of subgroup-specific treatment effects to leverage EC data while ensuring that the subgroup analysis is coherent with a primary analysis using RCT data alone. Our approach modifies subgroup-specific treatment effect estimators obtained through popular methods (e.g., linear regression) applied jointly to the RCT and EC datasets. We shrink these estimates so that their weighted average is close to a robust estimate of the average treatment effect in the overall trial population based on the RCT data alone. We study the proposed harmonized estimators with analytic results and simulations, and investigate standard performance metrics. The method is illustrated with a case study in glioblastoma.
READ FULL TEXT
