:: Workshop – Bayesianism II: Foundations and Applications ::

Tutorial 2 (thursday – 9h00 and friday – 9h00)
Bayesian Subgroup Meta-Analysis.
Lurdes Inoue (University of Washington).

Abstract:
Sudden cardiac death (SCD), usually due to a ventricular tachyarrhythmia, is the most common cause of death in the United States accounting for up to 350,000 deaths per year. Recent clinical trials have shown that the implantable cardioverter defibrillator (ICD), a device that monitors heart rhythms and delivers shocks if dangerous rhythms are detected, reduces the risk of SCD in certain patient populations. Given the substantial cost associated with ICD implantation and followup, the clinical and policy community currently seek to identify patients at risk of SCD and to improve the utilization of ICD implantation. With this goal, most investigations rely on stratified analysis of subgroups of patients defined by prognostic variables. However, these analyses are often exploratory and as such lack power to detect treatment effect in the subgroups. In this article, utilizing data from eight randomized controlled ICD clinical trials and Bayesian hierarchical models, we investigate whether the cumulative data provide evidence of treatment effect on mortality in the overall patient population and in the subgroups of interest.

Promotion and Support:
			CAPES-Proex Pós–graduação em Estatística-USP