|
Activity Number:
|
75
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Sunday, August 6, 2006 : 4:00 PM to 5:50 PM
|
|
Sponsor:
|
Biopharmaceutical Section
|
| Abstract - #306484 |
|
Title:
|
Survival Analysis on Recurrent Event Data: an Application to Alcoholism Study
|
|
Author(s):
|
Jian Han*+
|
|
Companies:
|
Bristol-Myers Squibb Company
|
|
Address:
|
5 Research Parkway, Wallingford, CT, 06492,
|
|
Keywords:
|
multivariate survival ; recurrent event
|
|
Abstract:
|
In alcoholism treatment clinical trials, one of conventional analyses of efficacy outcomes focuses on time to the first event ('any drinking' or 'heavy drinking'). The naturalistic course of drinking behavior during treatment often shows a gradual change in drinking before occurrence of a more stable drinking or abstinence pattern. Therefore, analyzing time to first event may not be sensitive enough to capture change in patient's drinking behavior if the treatment effect is observed in later relapse events. The multivariate survival time analysis or recurrent event analysis appears to be a better approach. Three commonly used strategies for multivariate survival time are available with emphasis of different aspects of recurrent events. These methods are explored and evaluated with respect to model assumption and validity in application in double blinded clinical trial data.
|
