SEM-based meta-analysis and MASEM

Research area: HTA-methods & steering instruments

Duration: May 2017 – November 2017

Language: German (with english summary)

Background:
Meta-analyses play an important role in synthesizing evidence on medical procedures. Usually, a single (univariate) meta-analysis for each individual outcome is calculated. This leads to the so called “multiple comparison problem”, since the risk of false positives is inflated and the reported P values lose their informative value. Additionally, interactions between the variables cannot be evaluated with univariate methods. This leads to the problem that complex relationships cannot be assessed. Recently, several articles have been published that used structural equation modeling (SEM) based meta-analysis and Meta-analytic Structural Equation Modeling (MASEM). SEM-based meta-analysis and MASEM promise to be a solution to the stated problems.

Aim:
The PhD project at hand aims to summarize the current research on these methods. Additionally, SEM-based meta-analysis and MASEM will be used in projects related to HTA. The focus of MASEM and SEM-based meta-analysis in these projects lies on showing differences of the results achieved by SEM-based meta-analysis and MASEM, compared to conventional univariate meta-analyses. Further, the use of MASEM and SEM-based meta-analysis for the evaluation of complex hypotheses should be shown.

Methods:
This PhD project will be supervised by assoc. Prof. Martin Voracek, D.Sc., D.M.Sc., Ph.D., M.Sc., M.Phil. (University of Vienna, Faculty of Psychology - Department of Basic Psychological Research and Research Methods). Three articles will be published with the following topics (may be subject to change):

1. A systematic review on recently published articles that used MASEM and SEM-based meta-analysis. As well as articles that described the method.
2. A comparison between the results achieved with univariate meta-analysis to MASEM or SEM-based meta-analysis for the assessment of a medical procedure.
3. Use of SEM-based meta-analysis or MASEM to measure the effect of conflict of interests on the results of randomized controlled trials.

Time Schedule:
From October 2016: Project planning
December 2016 – June 2018: Systematic Review on MASEM (Publication 1)
October 2017: Public presentation of the project plant at the University of Vienna
From November 2017: Work on projects in which MASEM is practically applied (Publication 2 and 3)