A guide to interpreting systematic reviews and meta-analyses in neurosurgery and surgery.

Abstract

Systematic reviews (SRs) and meta-analyses (MAs) are methods of data analysis used to synthesize information presented in multiple publications on the same topic. A thorough understanding of the steps involved in conducting this type of research and approaches to data analysis is critical for appropriate understanding, interpretation, and application of the findings of these reviews. We reviewed reference texts in clinical neuroepidemiology, neurostatistics and research methods and other previously related articles on meta-analyses (MAs) in surgery. Based on existing theories and models and our cumulative years of expertise in conducting MAs, we have synthesized and presented a detailed pragmatic approach to interpreting MAs in Neurosurgery. Herein we have briefly defined SRs sand MAs and related terminologies, succinctly outlined the essential steps to conduct and critically appraise SRs and MAs. A practical approach to interpreting MAs for neurosurgeons is described in details. Based on summary outcome measures, we have used hypothetical examples to illustrate the Interpretation of the three commonest types of MAs in neurosurgery: MAs of Binary Outcome Measures (Pairwise MAs), MAs of proportions and MAs of Continuous Variables. Furthermore, we have elucidated on the concepts of heterogeneity, modeling, certainty, and bias essential for the robust and transparent interpretation of MAs. The basics for the Interpretation of Forest plots, the preferred graphical display of data in MAs are summarized. Additionally, a condensation of the assessment of the overall quality of methodology and reporting of MA and the applicability of evidence to patient care is presented. There is a paucity of pragmatic guides to appraise MAs for surgeons who are non-statisticians. This article serves as a detailed guide for the interpretation of systematic reviews and meta-analyses with examples of applications for clinical neurosurgeons.