|
Activity Number:
|
453
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Wednesday, August 6, 2008 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Section on Bayesian Statistical Science
|
| Abstract - #301259 |
|
Title:
|
A Bayesian Bivariate Random-Effects Meta-Analysis for Two Correlated Outcomes Using Individual Patient Data
|
|
Author(s):
|
Ying Yang+ and Kao-Tai Tsai*+
|
|
Companies:
|
Bristol-Myers Squibb Company and Bristol-Myers Squibb Company
|
|
Address:
|
777 Scudders Mill Road, Plainsboro, NJ, 08536, 34 Baldwin Dr., Berkeley Heights, NJ, 07922,
|
|
Keywords:
|
multiple outcomes ; Bayesian ; bivariate meta-analysis
|
|
Abstract:
|
Multiple outcomes are often of interest in psychological clinical trials. For example, the Hamilton Depression Rating Scale (HAMD) and the Montgomery and Asberg Depression Rating Scale (MADRS) are two important outcomes. A separate univariate analysis is often used to evaluate treatment effect for each outcome independently. This approach ignores the correlation between outcomes. To fully assess treatment effect and understand the association between treatments and outcomes, we propose an approach which is a single Bayesian bivariate meta-analysis model that can jointly synthesize those endpoints and utilize their correlation. The individual patient data from five double-blinded, placebo-controlled trials using selegiline transdermal system for major depressive disorder is used to demonstrate the method.
|
