Lu Tian, PhD
Assistant Professor, Department of Health Research and Policy Stanford University School of Medicine
META ANALYSIS WITH SMALL NUMBER OF STUDIES
Wednesday, 13 Sept 2017, 3pm followed by social hour Room 2700, MissionHall: Global Health and Clinical Sciences Building, 550 16th Street, San Francisco, CA
Host: Mi-Ok Kim, PhD, (MiOk.Kim@ucsf.edu)
Meta analysis is a popular tool to synthesize evidence from multiple sources. In the presence of study heterogeneity, the random effects model is the commonly used statistical model for meta analysis. There are two major limitations for the statistical inference based on the random effects model: the dependence on the restricted parametric assumption and the asymptotic justification requiring large number of studies. When the number of studies is moderate or large, some exact inference procedures have been developed. However, when the number of study is less than 6, current methods are either inapplicable or invalid. We feel that if only a very small number of studies are available for meta analysis, then appropriate parametric assumptions are unavoidable in order to effectively summarize the data. In this talk we propose an exact inference procedure for parametric random effects model, which is valid regardless of the number of studies. We will illustrate the new proposal with both simulation study and real data example.