|
Activity Number:
|
198
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Monday, July 30, 2007 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Biopharmaceutical Section
|
| Abstract - #309431 |
|
Title:
|
Some Remarks on Multiple Imputations in Longitudinal Data Context
|
|
Author(s):
|
Mohammed K. Alam*+ and Marepalli B. Rao and Ramesh N. Amatya and Claudia Lara
|
|
Companies:
|
Kendle International Inc. and University of Cincinnati and Kendle International Inc. and Kendle International Inc.
|
|
Address:
|
1200 Carew Tower, Cincinnati, OH, 45202,
|
|
Keywords:
|
Multiple Imputations ; Longitudinal Data ; Clinical Trials ; Missing Data ; Inference ; Dropout
|
|
Abstract:
|
The focus is on Longitudinal Data Collection in Clinical Trials in which each patient in the sample is observed at some specified n time points with respect to a well-defined response measure. For each patient in the sample, either we have a complete record of the data at all time points of interest or some data are missing at some time points. This scenario includes dropouts, who drop out of the study completely after a certain time point. Molenberghs et al (2004) have focused on analysis of data consisting of complete records and dropouts. Our scenario is more general. There are two predominant reasons in imputing missing data points in our scenario. One is to have a complete record of responses for each patient in the study. Other reason is to examine the effect of imputation on inference. Some of the issues in this context will be discussed.
|
