Network meta-analysis: an introduction for clinicians

Abstract

Network meta-analysis is a technique for comparing multiple treatments simultaneously in a single analysis by combining direct and indirect evidence within a network of randomized controlled trials. Network meta-analysis may assist assessing the comparative effectiveness of different treatments regularly used in clinical practice and, therefore, has become attractive among clinicians. However, if proper caution is not taken in conducting and interpreting network meta-analysis, inferences might be biased. The aim of this paper is to illustrate the process of network meta-analysis with the aid of a working example on first-line medical treatment for primary open-angle glaucoma. We discuss the key assumption of network meta-analysis, as well as the unique considerations for developing appropriate research questions, conducting the literature search, abstracting data, performing qualitative and quantitative synthesis, presenting results, drawing conclusions, and reporting the findings in a network meta-analysis.

Keywords: Comparative effectiveness; Multiple treatment meta-analysis; Network meta-analysis; Transitivity.

Fig. 1

Network graph. Each node represents one drug, color-coded by class. Here, the size of the node is proportional to the number of participants randomized to that drug, and the width of the edge is proportional to the number of trials comparing two drugs [8]

Fig. 2

League table representing summary estimates from network meta-analysis [8]

Fig. 3

Ranking probabilities for any drug at any position [8]

Fig. 4

Cumulative ranking probabilities for each treatment in the network. The surface under the cumulative ranking curve (SUCRA) value is the probability each treatment has of being among the best of those in the network, with larger values representing higher ranking probabilities. In this figure, the top three treatments are bimatoprost, latanoprost, and travoprost, with SUCRA values of 99.6%, 86.56%, and 85.76%, respectively [8]