What are network meta-analyses (NMAs) and why are they used?
Indirect and mixed treatment comparisons are increasingly conducted when there is no direct evidence linking two interventions (A and B), when there is insufficient direct evidence or when there are more than two treatments.
Indirect treatment comparisons (ITCs) have no head to head evidence, and studies are linked through a third or more studies (C).
Networks which include head to head comparisons as well as indirect evidence are usually labelled mixed treatment comparisons (MTCs).
How do you start a network meta-analysis?
Depending on the condition being investigated and the number of trials available, network meta-analyses can become highly complex.
Robust network meta-analysis is based on a series of assumptions:
- that all relevant studies are included
- that individual studies are not biased
- that studies are homogeneous in terms of patient characteristics and study design
- the effect modifiers are absent or accounted for in the analysis
- that the studies are similar when considering an ITC or MTC
- that the studies (in an MTC) are consistent – that the direct and indirect evidence agree
Applications of network meta-analyses
Prepared by the Cochrane Editorial Unit and Cochrane Comparing Multiple Interventions Methods Group, this page contains information on the background to network meta-analysis, as well as some recent Cochrane reviews using NMAs to provide vital information for “real life” use in a variety of fields.