1. Network Meta-Analysis Techniques for Synthesizing Prevention Science Evidence
- Author
-
Georgios Seitidis, Emily A. Hennessy, Emily E. Tanner-Smith, Dimitris Mavridis, and Stavros Nikolakopoulos
- Subjects
Prevention science ,Hierarchy ,Consistency (database systems) ,Ranking ,Risk analysis (engineering) ,Computer science ,Meta-analysis ,Public Health, Environmental and Occupational Health ,Psychological intervention ,Pairwise comparison ,Field (computer science) - Abstract
Network meta-analysis is a popular statistical technique for synthesizing evidence from studies comparing multiple interventions. Benefits of network meta-analysis, over more traditional pairwise meta-analysis approaches, include evaluating efficacy/safety of interventions within a single framework, increased precision, comparing pairs of interventions that have never been directly compared in a trial, and providing a hierarchy of interventions in terms of their effectiveness. Network meta-analysis is relatively underutilized in prevention science. This paper therefore presents a primer of network meta-analysis for prevention scientists who wish to apply this method or to critically appraise evidence from publications using the method. We introduce the key concepts and assumptions of network meta-analysis, namely, transitivity and consistency, and demonstrate their applicability to the field of prevention science. We then illustrate the method using a network meta-analysis examining the comparative effectiveness of brief alcohol interventions for preventing hazardous drinking among college students. We provide data and code for all examples. Finally, we discuss considerations that are particularly relevant in network meta-analyses in the field of prevention, such as including non-randomized evidence.
- Published
- 2021