|
Activity Number:
|
149
|
|
Type:
|
Topic Contributed
|
|
Date/Time:
|
Monday, August 3, 2009 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Biopharmaceutical Section
|
| Abstract - #305795 |
|
Title:
|
Practical Statistical Considerations for Handling Dropouts in Longitudinal Clinical Trials
|
|
Author(s):
|
Cynthia M. DeSouza*+ and Abdul J. Sankoh
|
|
Companies:
|
Vertex Pharmaceuticals, Inc. and Vertex Pharmaceuticals, Inc.
|
|
Address:
|
130 Waverly Street, Cambridge, MA, 02139,
|
|
Keywords:
|
Dropout ; MNAR ; Selection Model ; MMRM ; MAR ; Shared Parameters Model
|
|
Abstract:
|
The presence of dropouts in longitudinal clinical trials could lead to complicated data analyses and potential uncertainty in the interpretation of trial results. We review inference strategies such as MMRM, ANCOVA-LOCF, weighted GEE, selection model and shared parameters model under both MAR and MNAR dropouts. We explore the sensitivity of these strategies to varying levels of dropout, varying intra-subject correlation, and different models for dropout by a simulation experiment, with emphasis on their practical differences. Implementation of these methods is conducted using SAS for MAR dropout, and WinBUGS with noninformative priors for MNAR dropout in order to handle estimation based on the joint likelihood function of parameters from both measurement and dropout processes.
|
- The address information is for the authors that have a + after their name.
- Authors who are presenting talks have a * after their name.
Back to the full JSM 2009 program |
