Objective: To assess the characteristics and core statistical methodology specific to network meta-analyses (NMAs) in clinical research articles.
Study Design and
What is new?Key findings Although the amount of evidence (the number of treatments and studies) included in published NMAs remains stable, the undertaking and reporting of statistical methods have significantly improved over the years. The assumptions underlying NMA are increasingly discussed and evaluated using appropriate methods. Less than 10% of NMAs published in 2014 and 2015 failed to evaluate the assumptions of the joint synthesis.
What this adds to what is knownThis meta-epidemiological study presents the largest collection of published NMAs over the past 16 years. It provides an overview of the structural characteristics and statistical methodology of 456 published networks of interventions. It shows that the statistical methods in NMA have considerably improved in all aspects and some, such as the use of appropriate methods to evaluate the plausibility of the assumptions, are now routinely performed. We conclude that the increasingly populous community of NMA methodologists is quickly advancing through the learning curve of statistical methods employed in NMA.
What is the implication, what should change nowThe updated description of the structural characteristics of the published NMAs can be used to inform pragmatic simulations studies and the development of methods that are relevant to the type of networks typically found in the medical literature. Future tutorials and training should be focused on improving the methodology and reporting on items that, although they have improved, their prevalence remains low, such as the formal exploration of heterogeneity and inconsistency and the presentation of all pairwise treatment effects.
Amisulpride is the only antipsychotic that outperformed placebo in the treatment of predominant negative symptoms, but there was a parallel reduction of depression. Cariprazine was better than risperidone in a large trial that was well-controlled for secondary negative symptoms, but the trial was sponsored by its manufacturer. Future trials should apply scientifically developed definitions such as the deficit syndrome and the persistent negative symptoms concept.
In network meta-analysis, it is important to assess the influence of the limitations or other characteristics of individual studies on the estimates obtained from the network. The percentage contribution matrix, which shows how much each direct treatment effect contributes to each treatment effect estimate from network meta-analysis, is crucial in this context. We use ideas from graph theory to derive the percentage that is contributed by each direct treatment effect. We start with the ‘projection’ matrix in a two-step network meta-analysis model, called the
H matrix, which is analogous to the hat matrix in a linear regression model. We develop a method to translate
H entries to percentage contributions based on the observation that the rows of
H can be interpreted as flow networks, where a stream is defined as the composition of a path and its associated flow. We present an algorithm that identifies the flow of evidence in each path and decomposes it into direct comparisons. To illustrate the methodology, we use two published networks of interventions. The first compares no treatment, quinolone antibiotics, non-quinolone antibiotics and antiseptics for underlying eardrum perforations and the second compares 14 antimanic drugs. We believe that this approach is a useful and novel addition to network meta-analysis methodology, which allows the consistent derivation of the percentage contributions of direct evidence from individual studies to network treatment effects.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.