B1475
Title: Sequential matched randomization to personalize randomization and improve covariate balance and trial efficiency
Authors: Jonathan Chipman - University of Utah (United States) [presenting]
Lindsay Mayberry - Vanderbilt University (United States)
Robert Greevy - Vanderbilt University (United States)
Abstract: Covariate-adjusted randomization (CAR) can reduce the risk of covariate imbalance and, when accounted for in the analysis, increase the power of a trial. Despite CAR advances, stratified randomization remains the most common CAR method. Matched randomization (MR) randomizes treatment assignment within optimally identified matched pairs based on covariates and a distance matrix. When participants enrol sequentially, sequentially matched randomization (SMR) randomizes within matches found "on the fly" to meet a pre-specified matching threshold. However, pre-specifying the ideal threshold can be challenging, and SMR yields less optimal matches than MR. Novel SMR extensions address these limitations and are studied in simplified settings and a real-world case study. SMR is compared to other CAR schemes, which highlights the different strengths of schemes. The case study provides an example in which adjusting for covariates in randomization (i.e., CAR with randomization-based inference) can be more powerful in testing the marginal average treatment effect than adjusting for covariates in a parametric model (i.e., complete randomization with ANOVA and ANCOVA).
