The PRISMA 2020 Checklist, Item by Item

Somewhere between the last extraction and the first draft, someone asks for the PRISMA checklist: the journal's submission form has a required upload, or a supervisor wants it attached to the thesis. If that is the first time you read its items, some of them will ask for numbers and decisions you no longer have.

Here is the short answer. The PRISMA 2020 statement carries the current PRISMA checklist for systematic reviews: 27 items naming what the finished report must contain, from the title to what you make publicly available. It is a reporting guideline, not a quality score. You fill it in at writing time, but it is cheapest to satisfy when you know the items before screening starts, because many of them describe things a review either records as it goes or never records at all.

The checklist has seven sections. The map below shows the items each section holds and what it asks for; the sections after it walk the items in plain words. Items with sub-items (10, 13, 16, 20, 23 and 24) are folded into one row each here; the official checklist splits them out.

What the PRISMA checklist is, and is not

PRISMA stands for Preferred Reporting Items for Systematic reviews and Meta-Analyses, and the expansion is the definition: it is a reporting guideline. The checklist tells you what the paper must disclose so a reader can follow and check every step. It does not tell you how to run the review, and it passes no judgement on your studies.

One checklist covers reviews with and without a meta-analysis. PRISMA 2020 replaced the 2009 version of the checklist; the FAQ below covers what that means for old templates, and how far the checklist reaches outside the fields it was written in.

Title, abstract and introduction: items 1 to 4

The first four items are the quick ones, usually settled in a single writing session.

Methods: items 5 to 15

This is the section that decides whether the review is reproducible, and most of its items describe decisions made months before writing.

Three of these items, 8, 9 and 11, ask the same pair of questions: how many people did this, and did they work independently. The answers are set when screening is set up; by writing time they are either in the project's records or gone. Item 7 has its own trap: it asks for the full strategy for every source, not one polished example with "similar terms were used elsewhere".

Results: items 16 to 22

The results items mirror the methods items: whatever you said you would assess, you now present, study by study and synthesis by synthesis.

Item 16b is the one that catches finished reviews. It asks you to cite the studies a reader would expect to see included and explain why each one is out. That is only possible if the reason was recorded at the moment of the full-text decision, not reconstructed from memory months later.

Discussion and other information: items 23 to 27

The last five items are the paper's accountability trail.

Note the honesty built into item 24: "we did not register" and "no protocol was prepared" are acceptable answers, silence is not. Registration itself is decided before screening; the protocol guide covers where and when.

The PRISMA 2020 for Abstracts checklist

Item 2 hands the abstract its own standard: a 12-item PRISMA abstract checklist for journal and conference abstracts, updated from the 2013 PRISMA for Abstracts statement. It is a compressed mirror of the main list.

Items 11 and 12 surprise people: yes, funding and the registration number belong in the abstract.

Which PRISMA checklist do you need?

PRISMA is a family of checklists, and the names sort by what you are reporting.

Two of these rows belong to other guides. The PRISMA-P checklist structures the protocol, and the protocol guide walks it with the registration routes. The PRISMA scoping review checklist, PRISMA-ScR, applies when the review maps evidence rather than answers one question; if you are not sure which review you are running, settle the type before the checklist.

Filling it in

The official page offers the checklist as a Word document or a PDF, and a web app completes it in the browser and exports either format. Fill the last column, where each item is reported, against the final manuscript, and upload the completed checklist with the submission.

The locations are suggestions, not rules. When word limits bite, the explanation and elaboration paper suggests pointing to the public protocol and moving detail to supplementary files; it is also the paper to open when an item's wording is unclear.

The items that hurt at writing time are the ones that depend on records kept during screening. In Littview, the flow diagram and the Methods draft are built from the review's own records, so the criteria, the screening setup and the counts are already logged by the time you open the checklist.

FAQ

Four questions come up most.

Is PRISMA a quality assessment tool? No. The checklist measures the completeness of the report, not the quality of the review or of its studies. Appraising the included studies is a separate step with its own instruments, item 11's named tool among them, and the appraisal guide maps which one fits which study design. If a marking rubric asks you to "appraise the paper with PRISMA", it is conflating the two.

Is there a separate PRISMA checklist for meta-analysis? No. The name covers both: Preferred Reporting Items for Systematic reviews and Meta-Analyses. A review that pools results answers the meta-analysis clauses inside items 13 and 20, the model used, the heterogeneity measures, the summary estimates; a review that synthesises without pooling answers the same items minus those clauses. There is no second checklist to find.

Which version is current, the PRISMA 2009 checklist or PRISMA 2020? PRISMA 2020, for every new review. The statement is explicit that the 2009 version should no longer be used, and the PRISMA 2020 abstract checklist now lives inside the main statement rather than as a separate one. Old templates circulate in course folders and departmental wikis, so download the checklist from prisma-statement.org rather than reusing an attachment.

Does the checklist apply outside health research? Yes. The statement was written primarily for reviews of health interventions, but its own scope note says the items apply to reviews of other interventions, social or educational ones for instance, and that many items apply to reviews asking other kinds of question entirely.

For this week: download the checklist before you touch a database. Put the items that describe process, 5 to 11 and 24, into the protocol while they are still decisions rather than memories, and record a reason for every full-text exclusion as you screen. Filled in that order, the checklist becomes what it was designed to be: a table of contents for work already done.