Two fundamental principles of evidence-based medicine are that decisions should be based on systematic summaries of the body of evidence, and that there is a hierarchy of evidence that arranges study designs by their susceptibility to bias. Hierarchies, however, are not absolute, and our current understanding of them has evolved into systems that integrate the hierarchy into more sophisticated structures for rating the quality of the body of evidence of specific health care questions. We have moved from rating the quality of individual studies to an "outcomes-centric" approach that rates the quality of evidence for each outcome across all available studies. Network meta-analysis is a sophisticated and promising technique that uses both direct and indirect study results to compare the relative effectiveness of multiple interventions on an outcome of interest. This technique may offer the best chance to understand the evidence when many competing treatments are available. Nevertheless, the evaluation of these networks requires careful considerations about the validity of the indirect comparisons, as well as other factors that may potentially affect the interpretation of the results. In particular, determinants of confidence related to incomplete reporting, inconsistency, and indirectness are of major concern in the analysis of network meta-analyses, and should be looked for and evaluated carefully when interpreting their results.