Meta-analysis is a popular statistical technique that synthesizes several trial results addressing the same research question. It results in more precise results compared to individual randomized clinical trials (RCTs). It allows exploring for heterogeneity (variation of effects across trials) mainly through subgroup and meta-regression analyses.
In most healthcare problems, there is a plethora of competing treatments and pairwise meta-analysis cannot estimate the comparative efficacy and safety of the competing treatments using all available evidence. Network meta-analysis (NMA) is an extension of pairwise meta-analysis when at least three competing treatments are available. When there is a network of interventions, information about each treatment comparison flows both directly, from studies including this comparison, and indirectly, from the rest of the studies. The benefits of NMA include estimating

• relative efficacy between all interventions, even those not compared directly in a RCT
• more precise effect estimates by synthesizing both direct and indirect information
• a hierarchy of interventions for each outcome.

As any statistical method, the validity of NMA results depends on the plausibility of the assumptions made. The main assumption of NMA is that of transitivity, implying that indirect estimates are valid and that the distribution of effect modifiers is similar across the various treatment comparisons. NMA of RCTs has been an established statistical method but new challenges have emerged.
In this talk, we will give a brief overview of pairwise and network meta-analysis, discuss their pros and cons, validity of assumptions, several misconceptions, challenges and future steps