Network meta-analysis and mixed treatment comparisons are statistical techniques to synthesize evidence based on various comparisons, some of which may be only indirect.
In many situations primary studies only compare two (or some other small number) of the conditions of interest. To bring all the information together network meta-analysis (NMA) was developed. For example some studies may compare A with B, some A with C but there may be no existing study comparing B with C. NMA provides a technology for doing this and providing indirect comparison between treatments.
An early source is Thomas Lumley's article Network meta--analysis for indirect treatment comparisons. There have also been useful discussion in a series of articles in the 2012 volume of the journal Research Synthesis Methods volume 3 starting at page 80
Software solutions are available in both Stata and R.
