To enhance interpretation of all presented graphical and numerical summaries we used three worked examples of NMAs. The first example compares 14 antimanic drugs for acute mania [14] (link). The network included 47 studies reporting on efficacy (measured as the number of responders out of total randomized) and 64 studies reporting on acceptability (measured as the number of dropouts out of total randomized). The second example is a network of 62 studies that evaluate the effectiveness of four different percutaneous coronary interventions for non-acute coronary artery disease [15] (link). The third example is a network of 27 studies forming a star-shaped network (i.e. all active treatments are compared only with placebo) that evaluated the effectiveness of six biologic agents for rheumatoid arthritis [16] (link). The outcome in this network was benefit from treatment defined as a 50% improvement in patient- and physician-reported criteria of the American College of Rheumatology (ACR50). The datasets and the STATA routines can be found online in www.mtm.uoi.gr and more detail is provided in the Appendix S1. To be able to carry out the analysis described below, version 3.01 (or later) of the command metan, version 2.6.1 (or later) of metareg and version 2.5.5 (or later) of mvmeta are required.
Free full text: Click here