Systematic Review

A protocol-driven, replicable synthesis

A systematic review is a synthesis method that answers a focused research question by following a pre-registered, publicly available protocol. All relevant studies are identified through comprehensive searches, selected by pre-defined eligibility criteria, appraised for methodological quality, and their findings are synthesized. Unlike narrative reviews based on subjective selection, systematic reviews minimize bias and ensure reproducibility through transparent, explicit methods. Reporting follows the PRISMA guideline.

Defining the Concept

A systematic review is a scientific method that answers a pre-specified research question by searching, selecting, quality-appraising, and synthesizing all relevant studies according to a pre-registered protocol. The word 'systematic' signifies that every stage is conducted and reported according to explicitly defined rules, which distinguishes it from narrative or traditional reviews whose conclusions can vary by author. Systematic reviews can stand alone as a qualitative synthesis or serve as the foundation for meta-analysis, a quantitative extension that statistically pools effect sizes across studies.

How It Works: Core Steps

A systematic review typically proceeds through six core stages. (1) The research question is structured using a PICO(S) framework and the protocol is registered in a repository such as PROSPERO. (2) A comprehensive search is conducted across multiple databases. (3) Studies are screened at title-abstract and full-text levels by two independent reviewers applying pre-defined eligibility criteria. (4) Methodological quality of included studies is assessed with domain-specific tools (e.g., RoB 2 for trials, QUADAS-2 for diagnostics). (5) Data are extracted and findings synthesized narratively or quantitatively. (6) Findings are reported using the PRISMA flow diagram and checklist.

A Concrete Example

Suppose a researcher wants to examine the effect of peer-teaching strategies on middle-school mathematics achievement. After registering a protocol in PROSPERO, systematic searches are run across ERIC, PsycINFO, and Web of Science. Two independent reviewers screen titles, abstracts, and full texts against pre-set eligibility criteria (randomised or quasi-experimental design, ages 10-15, presence of a control group). Risk of bias in each included study is assessed with RoB 2. The resulting evidence table supports both a narrative synthesis and a potential meta-analysis, while the PRISMA flow diagram transparently documents reasons for exclusion at each stage.

Common Pitfalls and Good Practice

Common pitfalls include failing to pre-register the protocol (which opens the door to selective reporting), omitting grey literature (theses, reports, conference proceedings) from searches, conducting quality appraisal with a single reviewer, and misapplying appraisal tools. Restricting searches to English-language publications also introduces language bias. Good practice requires pre-registering the protocol in PROSPERO, engaging at least two independent reviewers for screening and appraisal, selecting a quality tool matched to the study designs included, and reporting in full accordance with the PRISMA 2020 guideline.

Key terms

PRISMA: International guideline and checklist for transparent reporting of systematic reviews.
PROSPERO: International registry for prospective registration of systematic review protocols in health research.
PICO(S) Framework: Structured schema defining Population, Intervention, Comparison, Outcome, and Study design.
Risk of Bias: Assessment of the likelihood that included studies contain systematic error due to methodological weaknesses.
Grey Literature: Theses, reports, and conference papers not published through commercial academic publishers.