Network Meta-Analysis with Individual Participant Data

• Activity Number: 184
• Type: Contributed
• Date/Time: Monday, July 30, 2012 : 10:30 AM to 12:20 PM
• Sponsor: Section on Statistics in Epidemiology
• Author(s): Getachew Dagne*+ and Hendricks Crager and George Howe
• Companies: University of South Florida and University of Miami and The George Washington University
• Keywords: Meta-analysis ; mixed treatment comparison ; randomized trial ; statistical power ; network meta-analysis ; random effects

Abstract

Evidence-based prevention implementation requires a selection of the best among all competing interventions. Direct comparison of all relevant interventions for selecting the best preventive intervention is only possible within a randomized controlled trial involving such interventions. In the absence of this, combining information from various trials is highly desirable to arrive at sound conclusion on which an intervention of interest is most effective, and for which groups of subjects does the intervention work and under what situations. In this presentation, we show the power gains that may be obtained by combining trials while taking into account contextual settings, within and between-trial heterogeneity, and covariates. We illustrate our proposed methods using a simulation study.