Comparison of Power for ANCOVA Models Based on Central Tendency versus Nonparametric Analysis of Cumulative Responders for Clinical Trials with Continuous Response Endpoints

• Activity Number: 256
• Type: Contributed
• Date/Time: Monday, August 2, 2010 : 2:00 PM to 3:50 PM
• Sponsor: Biopharmaceutical Section
• Author(s): Kevin L. Lawson*+
• Companies: PPD, Inc.
• Keywords: Cumulative Proportion Responders Analysis ; Change From Baseline ; Wilcoxon Sum Rank Test ; ANCOVA ; CPRA ; CFB

Abstract

For clinical trials having a continuous response variable endpoint, the most prevalent primary analysis is a parametric analysis comparing the Change from Baseline (CFB) between two or more treatment groups using ANCOVA. A cumulative proportion of responders analysis (CPRA), where a responder is a patient with a CFB below a given value, can also be used to present results and test for significance. The purpose of this study was to explore the types of distributions of response variables from clinical studies with two treatment groups where the CPRA nonparametric analysis (e.g., Wilcoxon Sum Rank test) is more powerful than the ANCOVA analysis of mean CFB, and vice-versa. The Type I error and the power were compared for the nonparametric and parametric analyses using multiple samples drawn from data sets containing cholesterol reduction and pain intensity patients.