|
Activity Number:
|
145
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Monday, July 30, 2007 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Biopharmaceutical Section
|
| Abstract - #309004 |
|
Title:
|
Generalized Mixed Models for Noninferiority Trials
|
|
Author(s):
|
Robert Abugov*+
|
|
Companies:
|
Food and Drug Administration
|
|
Address:
|
7500 Standish Place, Rockville, MD, 20855,
|
|
Keywords:
|
noninferiority ; clinical trial ; positive control ; mixed effect model
|
|
Abstract:
|
Noninferiority trials consider a new therapy effective if the upper confidence bound on the difference in success between positive control and investigational treatment does not exceed a prespecified margin. Such trials commonly employ binary response data collected at multiple sites with multiple investigators per site, suggesting the generalized linear mixed effect model as appropriate for calculating the needed upper confidence bound on difference in success. This Monte Carlo study compares the realized accuracy of different methods for calculating confidence bounds by comparing actual coverage to intended percent confidence. If there is time, we will address model robustness as well as contingency planning should the most appropriate model fail to converge.
|
