Department of Biostatistics, UCLA School of Public Health
Title: Meta-Analysis for Observational Data
Abstract:
Meta-analysis is typically performed on randomized data, and standard meta-analysis models are better suited for investigating the performance of a randomized treatment. In contrast, we have data extracted from published papers on a number of observed covariates that may predict bad outcomes after syncope (fainting). We perform a multivariate Meta-analysis of these observational data to construct priors for analysis of an observational data set to predict bad outcomes after syncope.