Estimation of Heterogeneity Parameters in Multivariate Meta-Analysis

• Activity Number: 304
• Type: Contributed
• Date/Time: Tuesday, August 6, 2013 : 8:30 AM to 10:20 AM
• Sponsor: Biometrics Section
• Author(s): Abera Wouhib*+
• Companies: NCHS/CDC
• Keywords: DerSimonian and Laird ; Heterogeneity Parameter ; Hybrid Method ; Sidik and Jonkman

Abstract

Multivariate meta-analysis of random-effects model has the potential for estimating between-study covariance parameters whereas its univariate counterpart is limited only to estimating between-study variances. The most common challenge for univariate or multivariate meta-analysis is estimating heterogeneity parameters in non-negative domains under a random-effects model assumption. In this paper, two new heterogeneity parameters estimation methods in a multivariate setting are introduced; first by extending the Sidik and Jonkman univariate method of estimating heterogeneity parameter matrix to a multivariate model, and second by considering an iterative version of the Sidik and Jonkman method, namely a Hybrid method. These two methods are compared with extended DerSimonian and Laird methods by means of an illustrative example, and a simulation study. Finally, the methods are evaluated with respect to their bias in the simulation. Also, limitations of the heterogeneity parameters estimation methods applied in the multivariate meta-analysis case will be discussed.