A systematic review follows a rigorous, pre-planned methodology to minimise bias and synthesise all available evidence on a specific research question.
Critical appraisal of included studies using validated tools (Cochrane RoB, CASP, JBI, MMAT).
Formulated using PICO/PECO framework (Population, Intervention, Comparison, Outcome).
Transparent criteria for selecting studies based on design, population, outcomes, and language.
Systematic search across multiple databases with documented Boolean operators.
A systematic review identifies, selects, appraises, and synthesises all high-quality research evidence relevant to a clearly formulated question. It follows a reproducible methodology to minimise bias and provide reliable findings for evidence-based decision-making.
Rigorous protocol (PRISMA), exhaustive search, quality assessment, reproducible methodology, meta-analysis possible.
No protocol, selective citation, no formal quality assessment, author's perspective emphasised, narrative summary only.
Follow PRISMA 2020 guidelines for complete and transparent reporting
Structured summary with background, objectives, methods (databases, eligibility), results (studies included, main findings), and conclusions.
Establish the topic's importance. Explain why a systematic review is needed. State the research question using PICO/PECO format.
Describe protocol registration (PROSPERO), search strategy (databases, dates, terms), eligibility criteria, and appraisal tools.
PRISMA flow diagram showing identification, screening, eligibility, and inclusion. Table of study characteristics. Risk of bias assessment.
Synthesise main findings. Compare with other reviews. Explain heterogeneity. Discuss limitations of included studies and review process.
State overall strength of evidence. Provide specific recommendations for practice and future research directions.
Critical resources for conducting and reporting systematic reviews
27-item checklist and flow diagram for transparent reporting. Required by most journals for systematic reviews.
Register review protocols before data extraction. Prevents duplication and reduces reporting bias.
Assess methodological quality of randomised trials (RoB 2) and non-randomised studies (ROBINS-I).
Statistical synthesis of quantitative results. RevMan free from Cochrane. R packages: meta, metafor.