Testing for Publication Bias in Multivariate Random-Effects Meta-Analysis

• Activity Number: 59
• Type: Topic Contributed
• Date/Time: Sunday, August 9, 2015 : 4:00 PM to 5:50 PM
• Sponsor: Health Policy Statistics Section
• Author(s): Yong Chen* and Chuan Hong and Haitao Chu
• Companies: The University of Texas School of Public Health and The University of Texas School of Public Health and University of Minnesota, Twin Cities
• Keywords: Composite likelihood ; Outcome reporting bias ; Publication bias ; Score test

Abstract

Publication bias occurs when the publication of research results depends not only on the quality of the research but also on its nature and direction. The consequence is that published studies may not be truly representative of all valid studies undertaken, and this bias may threaten the validity of systematic reviews and meta-analyses - on which evidence-based medicine increasingly relies. Multivariate meta-analysis has recently received increasing attention for its ability reducing potential bias and improving statistical efficiency by borrowing information across outcomes. However, detecting and accounting for publication bias are more challenging in multivariate meta-analysis setting because some studies may be completely unpublished whereas some studies may selectively report part of multiple outcomes. In this paper, we propose an omnibus score test for both scenarios simultaneously. The proposed test is shown to be more powerful than the existing tests through simulation studies. In addition, the proposed test does not require estimating within-study correlations, which are often unknown or hard to obtain. The proposed test is illustrated through two meta-analysis examples.