|
Activity Number:
|
468
- Statistical Methods in Clinical Trials
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Wednesday, July 31, 2019 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Biopharmaceutical Section
|
|
Abstract #306752
|
Presentation
|
|
Title:
|
Percent Change from Baseline as an Endpoint in Clinical Trials
|
|
Author(s):
|
Jitendra Ganju* and Kefei Zhou
|
|
Companies:
|
Ganju Clinical Trials, LLC and Jazz Pharma
|
|
Keywords:
|
variance;
endpoint;
sample size
|
|
Abstract:
|
A well-known limitation of percent change from baseline as an endpoint is that its distribution is unstable if the moments do not exist. Despite this awareness there are still situations where the choice of percent change is risky unless there is a high degree of confidence in its stability. For example, an improvement defined as a lower score than baseline of a certain magnitude has a more stable distribution than one defined as a higher score of the same magnitude. Another example has to do with trial size planning: the variance of the estimated variance may be very large making power calculations unreliable. Implications for subgroup analyses will be discussed. An important practical rule developed by Marsaglia, which provides firm support for use of percent change, will be reviewed.
|
Authors who are presenting talks have a * after their name.
