The PRISMA 2020 reporting guideline is the updated Preferred Reporting Items for Systematic Reviews and Meta-Analyses statement: a 27-item checklist and a flow diagram that tell authors exactly what to report so that a reader can judge the conduct and trust the findings of a review. It is a reporting standard, not a method for doing the review, and it covers everything from the title and abstract through the search, selection, synthesis, and certainty of evidence.
What changed from the original statement
PRISMA 2020 replaced the earlier 2009 version to reflect how reviews are actually conducted today. The update expanded reporting of the full search strategy for every database, added items on how the review was registered and where the protocol lives, asked authors to describe automation tools and methods for assessing risk of bias, and broadened the synthesis items so that non-pooled syntheses are reported as transparently as a meta-analysis. The flow diagram was also redrawn to handle records found through databases separately from those found by other methods.
How the checklist is organised
Title, abstract, and introduction
The first items ask you to identify the report as a systematic review in the title and to give a structured abstract. The introduction items cover the rationale and the objectives, ideally framed with a question framework such as PICO.
Methods
This is the largest block. It asks for the eligibility criteria, the information sources, the complete search for at least one database, the selection process, the data collection process, how risk of bias was judged, the effect measures, the synthesis methods, and how reporting bias and certainty of evidence were assessed. Naming where the protocol was registered belongs here too.
Results, discussion, and other information
The results items map to the flow diagram and to a study-by-study presentation of characteristics, risk of bias, and results, including a forest plot where a meta-analysis was performed. The discussion interprets the findings and limitations, and the final block reports registration, protocol access, funding, and competing interests.
A practical way to keep the methods block honest is to write it directly from the protocol rather than from memory of what you did. Because the protocol already fixed the eligibility rules, the databases, and the planned synthesis before any study was seen, lifting those statements into the methods section guarantees the report matches the registration. Where the two diverge, the divergence is itself a reportable item: PRISMA asks you to declare any difference between the registered protocol and the completed review, which is why a maintained amendment log turns an awkward disclosure into a single tidy paragraph.
The two parts of the flow diagram
The PRISMA 2020 flow diagram traces records from identification through screening to the final included studies, with every exclusion counted and full texts excluded for stated reasons. The redesigned version separates records identified through databases and registers from those found by other methods, such as citation chasing. Getting the counts to reconcile is a common sticking point, and it is worth checking them against your de-duplication log before submission.
How to use the checklist without losing marks
Editors increasingly ask for a completed checklist with page numbers against every item, and the practical failures are predictable. Working through it as a submission tool, rather than a formality, avoids the recurring traps:
- Map each item to a page, not a section. “See methods” is not enough; reviewers want the exact location, and a vague pointer reads as an item half-addressed.
- Report the full search for at least one database verbatim. A summarised strategy fails the item; the actual line-by-line query belongs in a supplement, as covered in documenting and reporting your search.
- State the synthesis method even when you did not pool. A structured narrative synthesis must be described as deliberately as a meta-analysis, not left implied.
- Report certainty per outcome. A single overall sentence does not satisfy the item; each outcome needs its own GRADE rating.
- Disclose registration and protocol access. Cite the registration number and say where the protocol lives, an item authors routinely forget.
How PRISMA differs from a conduct guideline like Cochrane
A frequent source of confusion is treating PRISMA as instructions for doing a review. It is not. PRISMA tells you what to report once the work is done, while conduct guidance, such as the Cochrane Handbook, tells you how to do the work in the first place. A review can follow PRISMA reporting impeccably and still be poorly conducted, because complete reporting of a thin search merely makes the thinness visible. The two layers are complementary: sound conduct decides whether the review is trustworthy, and PRISMA reporting decides whether a reader can tell. This is the same separation that distinguishes a risk of bias judgement from a quality score, where the question is what the result actually means rather than how tidily it is presented.
Extensions for other review types
PRISMA 2020 is the core statement, but a family of extensions adapts it for specific designs. The best known is PRISMA-ScR for scoping reviews, PRISMA-P for review protocols, and others cover search reporting and abstracts. Whichever applies, the principle is the same: report enough detail that another team could audit and, in theory, repeat the review. That transparency is what separates a systematic review from a traditional literature review. If you want the diagram side of the standard handled in detail, see our guide to the PRISMA flow diagram.
Choosing the right extension matters more than it first appears, because the core checklist asks questions that do not always fit a non-pooled or exploratory design. A scoping review, for example, maps breadth rather than estimating an effect, so several methods items about effect measures and certainty simply do not apply, and forcing the core checklist onto it produces awkward “not applicable” entries that a methods editor reads as confusion. Matching the report to the right standard, whether the core statement for a full review of effects or the scoping extension for a review that maps the literature, is the difference between a checklist that documents the work and one that fights it.