More Powerful Analyzes of Crossover Trials with Baseline Measurements

• Activity Number: 591
• Type: Topic Contributed
• Date/Time: Thursday, August 7, 2014 : 8:30 AM to 10:20 AM
• Sponsor: Biopharmaceutical Section
• Author(s): Devan V. Mehrotra*+
• Companies: Merck
• Keywords: ANCOVA ; baseline adjustment ; covariance structure ; covariate ; crossover ; incomplete block

Abstract

In many crossover trials, for each participant, a continuous response of interest (e.g., blood pressure) is measured before and after administration of the assigned treatment within each period. For any given pairwise comparison of treatment means, we show that the power for statistical inference is greatly influenced by (i) the "structure" of the variance-covariance matrix of the vector of within-subject responses, and (ii) how the baseline (i.e., pre-treatment) responses are accounted for in the analysis. For (ii), we compare a common change from baseline analysis (which we advise against) with analysis of covariance approaches using relevant functions of the baselines as either period-specific or period-invariant covariates. We provide theoretical arguments and simulation-based type I error rate and power properties to support specific analysis recommendations for complete block (e.g., 2 period, two treatment [2x2]) and incomplete block (e.g., 2x4) crossover designs. Data from two clinical trials are used to reinforce the main points.