Analyzing Actual and Percentage Change as an Endpoint with or without Baseline as a Covariate in Clinical Trials

• Activity Number: 121
• Type: Contributed
• Date/Time: Monday, August 9, 2004 : 12:00 PM to 1:50 PM
• Sponsor: Biopharmaceutical Section
• Author(s): Pandurang Kulkarni*+ and Yongming Qu
• Companies: Eli Lilly and Company and Eli Lilly and Company
• Keywords: change ; percent change ; endpoint ; clinical trials ; sample size

Abstract

In many clinical trials from various therapeutic areas, the efficacy of a treatment is assessed using percentage change from baseline as the response variable as opposed to the actual change. Percentage change may be more intuitive to interpret than the actual change. However, choice of endpoint can have important implications. We reported the findings about using percentage change vs. actual change in several clinical trials at the 2002 JSM. Using the actual change from baseline is not always optimal if the within-subject variation is greater than the between-subject variation. We explored the testing aspects of comparing therapies using (a) the actual value, (b) the actual value with baseline as a covariate, (c) change from baseline. Choice of endpoint influences variation, power, sample size, and interpretations, leading to important differences in understanding of treatment effects and the underlying disease mechanisms. These factors will be discussed and guidance will be provided, in the context of a regulatory environment, in order to better understand the issues and to help optimize the design of clinical trials.