How to perform and write a meta‐analysis

Abstract

Meta-analysis is a statistical tool used to combine the results of multiple studies examining similar clinical questions, giving a quantitative pooled result. This increases the

How to conduct and report guidelines and position, best practice and consensus statements

“…Guidelines require a thorough review of the literature. In many cases, this will entail a systematic review which should be conducted in line with approved methodologies [9]. It is important that the search criteria should include an assessment of the relevance of articles to contemporary practice and the quality of the conducted research.…”

Effects and side‐effects of tranexamic acid: they both matter

“…We can examine in further detail the methodology used by Tsan et al [10]. Anaesthesia has defined standards for the performing and reporting of meta-analyses [11,12] and identified critical steps that are recommended for consideration both to readers and to authors of meta-analyses alike. General guidance for a meta-analysis is given by the preferred reporting items for systematic reviews and meta-analyses (PRISMA) statement [13] and this guidance was also followed by Tsan et al [10].…”

How to perform and write a trial sequential analysis

“…Therefore, if trial sequential analysis is planned, its main assumptions should be described in a protocol before starting the study. It is preferable for the protocol, including details of the trial sequential analysis, to be peer reviewed and published to avoid the analysis being data driven [14]; as an alternative to a publication in a journal, a systematic review or meta‐analysis (including the plan for the trial sequential analysis) can be registered on the International Prospective Register of Systematic Reviews (PROSPERO) [19]. The trial sequential analysis section of the protocol should at least include anticipated intervention effect (that is considered to be clinically relevant); anticipated heterogeneity between studies; and proportion of outcomes in the control group [14].…”

“…The standardised mean difference, however, cannot be used for trial sequential analysis for statistical reasons [17]. Like in meta-analysis, the model used for trial sequential analysis can be either a fixed or a random effects model with the choice being dependent on the anticipated heterogeneity [19]. For the random effects model, there are three methods: DerSimonian-Laird; Sidik-Jonkman; and Biggerstaff-Tweedie [17], with the first being the most used [17,19].…”