PRISMA 2020: What Changed and How to Comply

By Angel Reyes Last updated April 11, 2026 3 min read

PRISMA 2020 — the Preferred Reporting Items for Systematic Reviews and Meta-Analyses — was published by Page and colleagues in BMJ and Systematic Reviews in March 2021. It replaced the 2009 statement with a revised 27-item checklist, an updated flow diagram, and an expanded Explanation and Elaboration document. If you are starting a review in 2026, PRISMA 2020 is the baseline your journal and your reviewers will expect. This article summarizes what changed and what it means for day-to-day reporting.

What actually changed

The 2020 update is not a cosmetic refresh. Three categories of change matter:

New items. PRISMA 2020 adds explicit items for certainty-of-evidence assessment (GRADE), reporting of all outcomes and subgroup analyses rather than only pre-specified ones, and presentation of a completed PRISMA-S search reporting extension.
Restructured checklist. The checklist grew from 27 items organized under 7 headings in 2009 to 27 items under the same high-level headings but with clearer sub-items and more granular reporting requirements.
Revised flow diagram. The flow diagram now handles reports retrieved from registers separately from database records, and more clearly distinguishes records from reports (a single study can generate multiple reports).

For the full methodology context, see our reporting standards page and the systematic review guide.

The items most reviewers miss

In peer review, a handful of PRISMA items are consistently under-reported. Tighten these up before you submit:

Item 7 — Information sources. Name every database with platform and date range searched. "Searched PubMed and Google Scholar" is not compliant.
Item 8 — Search strategy. Include the full search strings for at least one database as a supplementary file. PRISMA-S requires all strategies.
Item 11 — Risk of bias. State the tool, whether it was applied in duplicate, and how conflicts were resolved.
Item 15 — Certainty assessment. Report GRADE (or equivalent) for each outcome, not just the overall review.
Item 24 — Registration and protocol. State the PROSPERO ID or explain why the review was not registered.

How to actually comply

The easiest way to comply is mechanical: work through the checklist as you write.

Download the checklist from the PRISMA website and paste it into your working document as a compliance table.
For each item, write the manuscript location (page and line numbers) where it is addressed.
When a draft is done, ask a co-author who was not involved in drafting to verify the mapping.
Include the completed checklist as a supplementary file — most journals require this.

The revised flow diagram

The 2020 flow diagram is more flexible than the 2009 version. Key points:

Two identification pathways. One for database and register records, one for records identified via other methods (citation searching, hand-searching, contacting authors).
Records vs reports distinction. Count unique studies at the end, not unique papers. A multi-paper study counts once.
Explicit exclusion reasons. At full-text stage, report counts by exclusion reason (wrong population, wrong design, etc.).

Grab our editable PRISMA flow diagram template and fill it in as you screen.

What PRISMA 2020 does not do

PRISMA is a reporting standard, not a methodology. It tells you what to report, not how to do the review. For the how, use the Cochrane Handbook for Systematic Reviews of Interventions (for healthcare reviews) or JBI methodology (for qualitative, prevalence, scoping, and mixed methods reviews).

PRISMA also does not replace protocol registration. Register on PROSPERO (or OSF for scoping reviews) before you start screening. The journal will ask.

A 10-minute compliance checklist

Before submission, verify:

[ ] Full PRISMA 2020 checklist attached as supplement, with page/line mappings
[ ] Search strategies for all databases in an appendix
[ ] Flow diagram uses the 2020 template, not the 2009 one
[ ] Risk of bias tool named and applied in duplicate
[ ] Certainty of evidence (GRADE) reported per outcome
[ ] PROSPERO ID or registration statement in methods
[ ] Deviations from protocol listed explicitly
[ ] Funding and conflicts of interest reported

Compliance is not a formality. Reviewers increasingly treat PRISMA items as a rubric, and editors will reject methodologically strong reviews that report poorly.