|
Activity Number:
|
198
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Monday, July 30, 2007 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Biopharmaceutical Section
|
| Abstract - #309003 |
|
Title:
|
Handling Missing Data in Long-Term Clinical Trials
|
|
Author(s):
|
Wei Deng*+ and Xuesong Guan and Chien-Feng Chen
|
|
Companies:
|
Amylin Pharmaceuticals, Inc. and Amylin Pharmaceuticals, Inc and Amylin Pharmaceuticals, Inc.
|
|
Address:
|
9360 Towne Centre Drive, San Diego, CA, 92121,
|
|
Keywords:
|
longitudinal trials ; last observation carried forward ; missing data ; mixed model repeated measure ; multiple imputation ; conditional model
|
|
Abstract:
|
This talk was motivated by a completed long-term clinical trial. The primary outcome was observed at different visiting schedules for different patients over three years. Our main interest is to estimate the change in primary outcome from baseline at the end of the trial. About one third patients withdrew from the trial prematurely. The impacts of missing data on inferences depend on the missingness process. Ignoring the missingness process may induce bias and inefficiency in the estimates. Thus the underlying missingness mechanism is explored and a few statistical models are considered, including last observation carried forward (LOCF), mixed model repeated measure (MMRM), multiple imputation (MI) and conditional model. A simulation study similar to the trial is also performed to compare these models.
|
