|
Activity Number:
|
276
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Tuesday, August 5, 2008 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Biopharmaceutical Section
|
| Abstract - #302612 |
|
Title:
|
On the Generalization of Closed Testing Procedures with Null Hypotheses Grouped into Families
|
|
Author(s):
|
Hanjoo Kim*+ and A. Richard Entsuah and Justine Shults
|
|
Companies:
|
University of Pennslyvania School of Medicine and Merck Research Laboratories and University of Pennsylvania
|
|
Address:
|
Department of Biostatistics, 423 Guardian Drive, Philadelphia, PA, 19104,
|
|
Keywords:
|
Closure method ; Multiple testing procedure ; Familywise error rate
|
|
Abstract:
|
We present a method of devising sequentially rejective multiple testing procedures with strong control of the familywise error rate for testing null hypotheses that are grouped into families. The method is a generalization of the closure method which provides a unified framework for constructing closed testing procedures in numerous applications. In particular, the method is applied to gatekeeping problems, in which hypotheses in a family can be tested only if hypotheses in the previous family have been rejected, to derive a class of stepwise gatekeeping procedures that are highly useful in clinical trials. We illustrate the computation of the adjusted p-values using a real clinical trial example which compares venlafaxine versus lithium monotherapies in patients with Bipolar II major depressive episode.
|
