|
Activity Number:
|
172
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Monday, August 3, 2009 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Biopharmaceutical Section
|
| Abstract - #305259 |
|
Title:
|
Comparison of Missing Data Approaches on Noninferiority Clinical Trials in Treatment of HIV Infection
|
|
Author(s):
|
Xia Xu*+
|
|
Companies:
|
Merck Research Laboratories
|
|
Address:
|
UG1C36, North Wales, PA, 19454,
|
|
Keywords:
|
missing data approach ; non-inferiority trial ; HIV infection
|
|
Abstract:
|
The proportion of subjects achieving HIV RNA levels below the assay limit at a specific time point is the primary endpoint in most HIV infection trials. The Noncompleter = Failure (NC=F) analysis is the most common missing data handling approach which is required by regulatory agencies and considered the most conservative approach. However, in non-inferiority trials, the NC=F analysis may result in bias and underestimate variability in favor of showing non-inferiority. Two other missing data approaches, Treatment Related Discontinuation = Failure and an Observed Failure approach, are often used as sensitivity analyses. Different missing data approaches are investigated and compared in the setting of non-inferiority trials by simulation.
|
