Introduction: Systematic reviews with network meta-analyses (NMAs) are complex techniques able to gather more evidence than conventional pairwise meta-analyses (1-3). The conduction, report, and publication of the most up-to-date and accurate information is paramount for evidence-based practices (4, 5). Our goal was to evaluate the proportion of NMAs following recommendations for the conduction and report of results, as well assess the publication delay.
Methods: A systematic review of NMAs on drug interventions was performed. PubMed and Scopus were searched without time-frame or language restrictions, and complemented by manual searches (updated 25th April 2017). The main characteristics of NMAs were collected by two independent reviewers including details of the systematic review process (protocol registration, databases searched, supply of search strategies and raw data). Delays in submission and publication were measured using the time from the last systematic search to the following dates: (i) article’s submission, (ii) acceptance, and (iii) online publication, measured in months. Descriptive analyses were conducted.
Results: We included 477 NMAs. Less than 60% followed recommendations for conducting and reporting the research (e.g. Preferred Reporting Items for Systematic Reviews and Meta-Analyses, PRISMA statement). Only 18% of studies have a published protocol or register number (e.g. International Prospective Register of Systematic Review, PROSPERO register). Around 9% of studies used only one electronic database for the systematic search and other 6% did not report this information. Medline (92.9%), Cochrane Library (78.6%) and Scopus (77.9%) were the most used databases. Only one third of NMAs (n=155) presented a replicable full-search strategy. Supplemental material with raw or additional data were supplied by 282 studies. Statistical analyses were conducted with 289 from the 477 NMAs providing information on publication dates. The median time from the least search to articles submission was around 6 months, while from last search to acceptance was 11 months and to final online publication was around a year. Only 11 NMAs performed updates of the systematic search after submission or acceptance, but before publication.
Conclusions: Efforts from authors in updating the systematic search before submission are needed to reduce the time of evidence production. Editors and producing agencies should ensure that guidelines and recommendations for NMAs conduction and report are strictly followed before publication.