|
Activity Number:
|
276
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Tuesday, August 5, 2008 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Biopharmaceutical Section
|
| Abstract - #300635 |
|
Title:
|
Type I Error Rate Adjustment for Superiority Test Conditional on Establishment of Noninferiority in Clinical Trials
|
|
Author(s):
|
Jiacheng Yuan*+ and Tiejun Tong
|
|
Companies:
|
Medtronic CardioVascular and University of Colorado, Boulder
|
|
Address:
|
3576 Unocal Place, Santa Rosa, CA, 95403,
|
|
Keywords:
|
non-inferiority ; superiority ; type I error rate adjustment ; conditional test ; clinical trial
|
|
Abstract:
|
In clinical trials, it is often desirable to test superiority conditional on establishment of noninferiority. Committee for Proprietary Medicinal Products issued guidelines accepting that no type I error rate adjustment is required for this switch. However, a researcher pointed out that switching between noninferiority and superiority without any adjustment reduces to simultaneous testing of both, which not only has a logic flaw similar to post-hoc specification of the null hypothesis but also leads to a higher chance for erroneous claiming superiority compared with only one null hypothesis to be tested. We suggest an adjustment which makes the second step test depend on the first step test, hence it is a conditional test in the real sense and the switch does not reduce to simultaneous testing. Also our suggested adjustment decreases the chance of erroneous claiming superiority.
|
