systematic review

systematic review
Title	Systematic review
Discipline	Research methods
First published	1979
Keywords	Evidence synthesis, Meta-analysis, PRISMA, Cochrane

systematic review

A systematic review is a structured, reproducible synthesis of research evidence that addresses a focused question using explicit methods. It complements primary studies and narrative reviews by applying pre-specified protocols to identify, select, appraise, and synthesize findings from multiple sources. Leading organizations and works that shaped the approach include Cochrane Collaboration, The Campbell Collaboration, Institute of Medicine (US), PRISMA Statement, and key contributors such as Archie Cochrane, Iain Chalmers, David Sackett.

Definition and scope

A systematic review is defined by rigorous protocol-driven procedures employed to collate empirical evidence that fits pre-determined eligibility criteria, distinguishable from narrative reviews, scoping reviews, and rapid reviews produced by bodies like World Health Organization, National Institute for Health and Care Excellence, and Agency for Healthcare Research and Quality. Its scope may address clinical effectiveness, diagnostic accuracy, prognostic factors, or policy interventions evaluated in trials registered in platforms such as ClinicalTrials.gov and ISRCTN Registry and indexed in bibliographic databases including MEDLINE, Embase, and Cochrane Central Register of Controlled Trials.

Methodology

Methodology centers on a priori protocol development often registered with repositories like PROSPERO or overseen by institutions such as Cochrane and Campbell. Core methodological components parallel standards recommended by the Institute of Medicine (US), PRISMA guidelines, and agency handbooks from World Health Organization and European Medicines Agency. Teams typically include content experts, methodologists, and information specialists affiliated with universities such as Johns Hopkins University, University of Oxford, Harvard University, or research centers like National Institutes of Health.

Search strategy and study selection

Search strategies use tailored strings across databases including MEDLINE, Embase, PsycINFO, CINAHL, and trial registries like ClinicalTrials.gov, with supplementary searches in gray literature from organizations such as World Bank and repositories like OpenGrey. Selection involves dual screening by reviewers trained in standards from Cochrane Handbook and PRISMA extensions, with conflicts resolved by arbitration often involving scholars from institutions like University of Toronto or Karolinska Institutet. Transparent selection yields a study flow documented similarly to diagrams popularized by PRISMA Statement and often reported alongside registration records in PROSPERO.

Quality assessment and risk of bias

Quality appraisal and risk-of-bias assessment employ structured tools developed by entities such as Cochrane Collaboration (RoB tools), GRADE Working Group (certainty of evidence), Newcastle-Ottawa Scale, and QUADAS-2 for diagnostic accuracy. Assessments probe selection, performance, detection, attrition, and reporting biases as conceptualized in landmark works associated with scholars like John Ioannidis and Doug Altman. Outcomes from these tools inform sensitivity analyses and judgments used by guideline developers at organizations like National Institute for Health and Care Excellence and World Health Organization.

Data synthesis and meta-analysis

Data synthesis ranges from narrative summary to quantitative meta-analysis using fixed-effect or random-effects models implemented in software from vendors such as Cochrane's RevMan, R Project for Statistical Computing packages (meta, metafor), and commercial platforms like StataCorp's Stata. Meta-analyses address heterogeneity using statistics attributed to researchers like Thomas R. Knapp and methods such as I² and tau², and explore publication bias with funnel plots inspired by work from Robert Egger and Douglas Altman. Advanced synthesis includes network meta-analysis shaped by contributions from groups at University of Bristol and London School of Hygiene & Tropical Medicine, and individual participant data meta-analysis coordinated by consortia like Early Breast Cancer Trialists' Collaborative Group.

Reporting standards and guidelines

Reporting follows checklists and standards promulgated by PRISMA Statement, Cochrane Handbook for Systematic Reviews of Interventions, and extensions for diagnostic, harms, and equity questions from groups linked to Equator Network. Protocols are commonly registered in PROSPERO or published in journals such as BMJ, The Lancet, and Cochrane Database of Systematic Reviews, while updates and living systematic reviews are practiced by teams at institutions like World Health Organization and McMaster University.

Applications and limitations

Systematic reviews underpin clinical practice guidelines from bodies such as National Institute for Health and Care Excellence, U.S. Preventive Services Task Force, and European Medicines Agency and inform policy at organizations like World Health Organization and UNICEF. Limitations include dependency on available primary studies, risk of publication bias reported by John Ioannidis, resource intensiveness noted by Cochrane Collaboration, and challenges synthesizing heterogeneous interventions highlighted in work by Iain Chalmers and Archie Cochrane. Emerging concerns involve reproducibility, conflicts of interest scrutinized in investigations linked to journals like JAMA and The BMJ, and methodological debates addressed at conferences hosted by Society for Research Synthesis Methodology.

Category:Research methodology