PRISMA 2020 Step-by-Step: How to Report a Systematic Review Correctly

PRISMA 2020 is the most widely used reporting guideline for systematic reviews. If your paper feels complete but reviewers still ask for missing details, PRISMA is usually the reason. This guide walks through PRISMA 2020 step by step, so you can write a systematic review that is transparent, reproducible, and publication-ready.

Recommended reading for your workflow: How to Build a PubMed Search Strategy, How to Critically Appraise a Study, How to Write a Research Protocol.

What PRISMA 2020 is and what it is not

PRISMA is a reporting guideline, not a quality score. It tells you what to report, not what design to choose. Even a strong review can look weak if reporting is incomplete.

Outbound reference: PRISMA Statement.

Step 1: Title and abstract (do not lose points early)

In the title, explicitly state this is a systematic review and, if applicable, a meta-analysis. In the abstract, follow structured reporting: background, objectives, eligibility, information sources, risk of bias, synthesis method, results, limitations, and conclusions.

Tip: Many journals expect PRISMA for abstracts too. Consider using PRISMA for Abstracts guidance when available.

Step 2: Methods section (the most important PRISMA area)

Reviewers judge credibility mainly from methods. PRISMA expects you to report:

• Eligibility criteria: population, study designs, outcomes, timeframe, language limits and why
• Information sources: which databases, registries, websites, reference checks
• Full search strategy: include the exact query for at least one database
• Selection process: how many reviewers, independent screening, conflict resolution
• Data collection: extraction fields, duplicate extraction, handling missing data
• Risk of bias assessment: which tool and how it was applied
• Effect measures and synthesis: models, heterogeneity, subgroup plans

To strengthen methods, align your plan with the Cochrane Handbook.

Outbound reference: Cochrane Handbook.

Step 3: PRISMA flow diagram (make it precise)

Your flow diagram should match your actual screening process. Common errors include mismatched counts, unclear duplicates removal, and missing reasons for full-text exclusions.

Best practice:

• Record duplicates removed and the tool used
• Separate title/abstract exclusions from full-text exclusions
• List the top reasons for full-text exclusion

If you are managing screening inside SciTrack, keep an audit trail so counts are automatic and consistent with PRISMA.

Internal link: Systematic Reviews Workspace.

Step 4: Results section (show what you found, not only what you think)

PRISMA expects results to include:

• Study selection: counts and flow diagram
• Study characteristics: tables with population, interventions, outcomes, follow-up
• Risk of bias results: per study and domain summaries
• Synthesis results: pooled estimates, confidence intervals, heterogeneity
• Reporting biases: funnel plots or assessments if appropriate

Do not hide heterogeneity. Explain it with subgroup or sensitivity analyses when possible.

Step 5: Discussion section (interpretation with discipline)

A strong discussion answers four questions:

1. What is the main finding and how strong is the evidence?
2. How does it compare with previous evidence?
3. What are the key limitations (risk of bias, heterogeneity, indirectness)?
4. What are the implications for practice and future research?

Be careful with causal language if your included evidence is observational.

Step 6: Transparency items reviewers look for

• Protocol registration (PROSPERO) or an explanation if not registered
• Availability of extraction forms and data when possible
• Funding and conflict of interest declarations

Outbound reference: PROSPERO.

Common PRISMA mistakes (and quick fixes)

• Mistake: search strategy not reproducible. Fix: paste full query and date searched.
• Mistake: unclear screening process. Fix: state number of reviewers and conflict resolution.
• Mistake: risk of bias unclear. Fix: specify tool and show domain results.
• Mistake: pooled estimate presented without context. Fix: interpret heterogeneity and applicability.

Conclusion

PRISMA 2020 helps you report systematic reviews clearly and transparently. Use it as a map: describe your search, selection, extraction, risk of bias, and synthesis in a reproducible way. When PRISMA is done well, peer review becomes easier and your conclusions become more trustworthy.

For more methods posts, explore the SciTrack Blog.