|
Activity Number:
|
79
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Sunday, August 2, 2009 : 4:00 PM to 5:50 PM
|
|
Sponsor:
|
Biopharmaceutical Section
|
| Abstract - #303451 |
|
Title:
|
Comparison of Multiple Testing Methods in a Trial with Co-Primary Endpoints
|
|
Author(s):
|
Moumita Sinha*+
|
|
Companies:
|
Boehringer Ingelheim Pharmaceuticals, Inc.
|
|
Address:
|
900 Ridgebury Road, Ridgefield, CO, 06877,
|
|
Keywords:
|
Multiple testing ; Multiple endpoints ; Multiple doses ; Minimum Effective dose (MED)
|
|
Abstract:
|
Four multiple testing methods have been compared for a clinical trial setting with co-primary endpoints. The methods are 1) Compute adjusted p-values for all dose comparisons and endpoints by Hochberg's method. 2) Hochberg's correction with a gate-keeping approach. 3) A correction of the above method to control the overall type 1 error rate. 4) An intersection union test (IUT) for the two endpoints at each dose, using a Bonferroni correction for the number of dose comparisons. Three dose-response curves are considered. The second method identifies the correct minimum effective dose (MED) with greatest probability. However the type 1 error rate is not controlled. Among the other methods, the IUT has higher power and identifies the correct MED with greatest probability in most situations that have been compared.
|
