Confidence Interval for Back-transformed Between-Treatment Mean Percent Change Difference Using Method of Variance Estimates Recovery

• Activity Number: 667
• Type: Contributed
• Date/Time: Thursday, August 4, 2011 : 10:30 AM to 12:20 PM
• Sponsor: Biopharmaceutical Section
• Author(s): Yu Chen*+ and Yue Shentu
• Companies: Merck & Co., Inc. and Merck & Co., Inc.
• Keywords: confidence interval ; method of variance estimation recovery ; log percent change ; mean difference

Abstract

Percent change from baseline is a common metric used in clinical trial analysis. The usual distributional assumptions for the analysis of covariance are often violated when the response is in the percent-change scale. The log transformation for the percent changes is frequently applied to circumvent this issue. When the difference between treatment in percent change from baseline is of interest, appropriate back-transformation needs to be carried out for the point estimate and confidence interval (CI). In this article, we propose a new method to construct the CI for between-treatment difference in mean percent changes using the method of variance estimates recovery (MOVER). The advantage of the proposed CI compared to other existing approaches is demonstrated by extensive simulations.