Open AccessIntroduction to network meta-analysis for evidence synthesis
Author(s) -
Catalina Gracia González-Xuriguera,
Laura Vergara-Merino,
Luis Garegnani,
Luis E. Ortiz-Muñoz,
Nicolás Meza
Publication year - 2021
Publication title -
medwave
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.178
H-Index - 7
ISSN - 0717-6384
DOI - 10.5867/medwave.2021.06.8315
Subject(s) - meta analysis , categorization , transitive relation , grading (engineering) , psychological intervention , ranking (information retrieval) , systematic review , computer science , psychology , medicine , medline , artificial intelligence , mathematics , civil engineering , combinatorics , psychiatry , law , political science , engineering
This article belongs to a collaborative methodological series of narrative reviews about biostatistics and clinical epidemiology. The goal is to present basics concepts concerning the systematics reviews of multiple treatments comparisons with network meta-analysis. For clinical ques-tions with several therapeutic alternatives to be compared, the central question is how to classify or rank their effectiveness (benefit and harm) to choose the best option. The network meta-analysis aims to answer questions related to the effectiveness and safety of comparing multiple treatments by the simultaneous analysis of results raised from direct and indirect comparisons. The network geometry is the general graphical representation of the network meta-analysis and allows to understand and assess the strength of comparisons. The network meta-analysis should check several assumptions to be valid, especially the transitivity assumption, which allows assuming that there are no systematic differences among the included comparisons, except their compared interventions. Thus, it is possible to know the relative therapeutic effectiveness of each pair of interventions included in the network meta-analysis and their ranking in terms of categorization. It has been proposed to use a modified Grading of Recommendation, Assessment, Development, and Evaluation (GRADE) approach considering the distinctive features of network meta-analysis to assess the certainty of the evidence for each comparison and the ranking of interventions.