Abstract
In this manuscript we use realistic data to conduct a network meta-analysis using a Bayesian approach to analysis. The purpose of this manuscript is to explain, in lay terms, how to interpret the output of such an analysis. Many readers are familiar with the forest plot as an approach to presenting the results of a pairwise meta-analysis. However when presented with the results of network meta-analysis, which often does not include the forest plot, the output and results can be difficult to understand. Further, one of the advantages of Bayesian network meta-analyses is in the novel outputs such as treatment rankings and the probability distributions are more commonly presented for network meta-analysis. Our goal here is to provide a tutorial for how to read the outcome of network meta-analysis rather than how to conduct or assess the risk of bias in a network meta-analysis.
Highlights
Frequently used as a synonym for network meta-analysis, a mixed treatment comparisons meta-analysis is a subset of a network meta-analysis which has ‘A statistical approach used to analyze a network of evidence with more than two interventions which are being compared indirectly, and at least one pair of interventions compared both directly and indirectly’ (Coleman, 2010)
We describe a Bayesian approach to network meta-analysis, as reviews using this approach often provide more outputs that require interpretation compared to a frequentist network meta-analysis
The aim of this paper was to describe in lay terms the interpretation of outputs frequently reported with network meta-analyses
Summary
Creativecommons.org/licenses/by/4.0/), which permits unrestricted re-use, distribution, and reproduction in any medium, provided the original work is properly cited. Network meta-analysis offers the advantage of enabling the combined assessment of more than two treatments, and the mixed treatment comparison ‘component’ of meta-analysis has the additional feature of enabling indirect estimation of treatment comparisons that might not be available in the literature in a formal statistical manner (Lu and Ades, 2004; Dias et al, 2014). One pairwise meta-analysis compared the efficacy of tulathromycin to florfenicol and the other tulathromycin to tilmicosin (Wellman and O’Connor, 2007). Each meta-analysis used only direct comparison of tulathromycin to florfenicol or tulathromycin to tilmicosin from the published literature This dual approach to pairwise meta-analysis, left readers without an estimate of the comparative efficacy of florfenicol to tilmycosin because no randomized controlled trials were available at the time for that direct effect. Given that many producers and veterinarians are interested in comparisons of interventions for which no randomized controlled trials are available, network meta-analysis is very useful and is increasingly being adopted
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
