|
Activity Number:
|
528
|
|
Type:
|
Topic Contributed
|
|
Date/Time:
|
Thursday, August 10, 2006 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Biopharmaceutical Section
|
| Abstract - #306977 |
|
Title:
|
Bootstrap Generalized Estimating Equation for Treatment Differences for Randomization Trials with Correlated Data
|
|
Author(s):
|
Duo Zhou*+ and Zengri Wang
|
|
Companies:
|
Medtronic, Inc. and Medtronic, Inc.
|
|
Address:
|
1015 Gramsie Road, MS Z240, Saint Paul, MN, 55126,
|
|
Keywords:
|
bootstrap ; correlated data ; generalized estimating equation ; multiple events
|
|
Abstract:
|
Generalized estimating equations are used commonly to estimate differences in treatment success rates differences in randomization trials with events occurring multiple times within patients. A naïve estimation of treatment effects without considering within-patient correlation may lead to biased estimates, especially when the event distribution is highly skewed. We consider a bootstrap-based estimating equation method for estimating treatment differences between groups. Results from several commonly used approaches are compared with respect to point estimates and confidence intervals. Simulation studies are carried out to assess the sensitivity of the methodology. An example involving a device trial with an implantable cardioverter defibrillator will be given to illustrate the details of calculation.
|
