|
Activity Number:
|
104
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Monday, August 4, 2008 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Section on Health Policy Statistics
|
| Abstract - #302158 |
|
Title:
|
Bayesian Models for the Meta-Analysis of Sparse Tables
|
|
Author(s):
|
Eloise Kaizar*+ and Joel Greenhouse and Howard Seltman
|
|
Companies:
|
The Ohio State University and Carnegie Mellon University and Carnegie Mellon University
|
|
Address:
|
Department of Statistics, Columbus, OH, 43210,
|
|
Keywords:
|
meta-analysis ; Bayesian models ; rare events ; sparse tables
|
|
Abstract:
|
Recent concern around the possible association of severe but rare adverse events with commonly-used medications (e.g., antidepressants, COX-2 inhibitors) has led to a number of meta-analyses of adverse events observed in randomized controlled trials. Because so few events are observed in each trial, many traditional meta-analysis methods are computationally unable to be applied in these analyses without ad-hoc adjustment. Bayesian models have been proposed to overcome many of the computational issues associated with rare event meta-analysis. However, the behavior of these models has not been adequately explored. We illustrate how different parameterizations and corresponding choice of prior specification impact the performance of Bayesian estimates and their robustness.
|
