PRISMA 2020 Guidelines: Complete Checklist for Systematic Reviews

Systematic reviews live or die on how transparently they explain what was actually done. Most medical and health-science journals now expect authors to work from the PRISMA 2020 checklist when writing one up, since it exists specifically to enforce that transparency. Its arrival in 2021 replaced a 2009 version that had gone more than a decade without an update, and Matthew Page’s working group at Monash University led the revision after surveying more than a hundred systematic review methodologists and journal editors about what the old checklist was missing.^1,3 A manuscript that still walks through the 27 items as they read in 2009 will already be out of date on several points, including how much of the search strategy has to be shown in full. What follows works through the current checklist section by section, the flow diagram that goes with it, a worked example of how one item should actually read in a manuscript, and the gaps that show up most often when journals check a submission against it.

In Simple Words: PRISMA 2020 is not a set of rules for how to run a systematic review. It is a checklist for how to write one up afterward, so a reader can see exactly what was searched, how studies were chosen, and how the results were combined. Twenty-seven items cover everything from the title page to the data-availability statement, and a four-part flow diagram shows how many records went in and how many studies came out the other end. Most journals in medicine and health research now ask for a completed PRISMA 2020 checklist alongside the manuscript, and missing or vague answers on it are one of the more common reasons a submission gets sent back before it reaches peer review.

What Is PRISMA 2020 and Who Actually Has to Follow It

PRISMA stands for Preferred Reporting Items for Systematic Reviews and Meta-Analyses. Despite the name, it doesn’t tell anyone how to design or conduct a review; it specifies what to report once the review is finished, so a reader can judge whether the methods support the conclusions.¹ That distinction trips people up constantly. Choosing a risk-of-bias tool or a meta-analysis model is a methods decision, governed by things like Cochrane guidance or the specific research question. Whether that choice gets described clearly enough for someone else to evaluate it is a reporting decision, and that’s the part PRISMA 2020 governs.

Authors of any systematic review, with or without meta-analysis, are the primary audience, but the checklist also functions as a working tool for peer reviewers and editors, who often check a submission against it item by item before deciding whether to send it out for full review. Cochrane reviews, health technology assessments, and clinical guideline developers commonly adopt it as well, sometimes with a topic-specific extension layered on top, such as PRISMA-ScR for scoping reviews or PRISMA-DTA for diagnostic test accuracy reviews. That core 27-item statement, published simultaneously in BMJ and several other journals in March 2021, is the version referenced throughout this guide.¹

What Changed Between PRISMA 2009 and PRISMA 2020

Six of the original 2009 items survived the update largely untouched, according to the survey the working group ran before drafting the new statement.³ Everything else got reconsidered, and a handful of changes matter enough that carrying forward a 2009-era habit will actively hurt a 2026 submission.

Search strategy reporting is the change authors miss most often. PRISMA 2009 asked for a full search strategy covering at least one database, with other databases described in less detail. PRISMA 2020 asks for the full strategy for every database, register, and website searched, not just one representative example.¹ A methods section built on the older habit, one detailed MEDLINE strategy plus a sentence noting that “similar terms were adapted for other databases,” no longer satisfies the current checklist.

Protocol registration is new territory entirely, absent from the 2009 statement altogether. PRISMA 2020 devotes three sub-items (24a to 24c) to registration and protocol reporting, reflecting how routine PROSPERO registration has become since the register launched in 2011.^1,7 Certainty assessment is the other major addition, and items 15 and 22 ask authors to describe the methods used to judge confidence in the body of evidence, commonly GRADE, and to report that certainty rating for each outcome. Neither existed as a distinct item in 2009.

A few smaller structural shifts round out the update. Synthesis methods split into six sub-items (13a to 13f), separating the decision of which studies belong in which synthesis from the mechanics of running it. Results of syntheses gained a parallel structure (20a to 20d) for the same reason, and the flow diagram itself changed shape, distinguishing records identified through databases and registers from records found through other methods such as citation searching, and separating “reports not retrieved” from “reports excluded” as two different failure points in the pipeline.

The PRISMA 2020 Checklist: All 27 Items, Section by Section

Below, a table summarizes what each item asks for, written in plain language rather than reproduced verbatim from the BMJ publication. The official checklist and its fillable template are freely available from the PRISMA statement website.

Item	Section	What to Report
1	Title	Identify the report as a systematic review, noting whether it includes a meta-analysis.
2	Abstract	Provide a structured summary following the separate PRISMA 2020 for Abstracts checklist.
3	Rationale	Explain why the review was needed given what was already known.
4	Objectives	State the review question explicitly, ideally framed with PICO.
5	Eligibility criteria	Define inclusion and exclusion criteria and how studies were grouped for each synthesis.
6	Information sources	List every database, register, website, organization, and reference list searched, with the date each was last checked.
7	Search strategy	Give the full search strategy for every database, register, and website used, including filters and limits.
8	Selection process	State how many reviewers screened each record and report, whether they worked independently, and any automation tools used.
9	Data collection process	Describe how data were extracted, by how many reviewers, and how disagreements or missing data were handled.
10a-10b	Data items	List every outcome and every other variable, participant characteristics, funding source, and so on, that the review sought data on.
11	Risk of bias assessment	Name the tool used, which domains were assessed, and how many reviewers judged each study.
12	Effect measures	Specify the statistic used for each outcome, such as risk ratio, mean difference, or hazard ratio.
13a-13f	Synthesis methods	Describe how studies were grouped, how data were prepared, how results were tabulated or plotted, the synthesis method and software, how heterogeneity was explored, and any sensitivity analyses.
14	Reporting bias assessment	Describe methods used to assess bias from missing results, such as funnel plots or Egger’s test.
15	Certainty assessment	Describe the methods used to judge certainty in the evidence for each outcome, if this was done.
16a-16b	Study selection	Report the numbers at each stage of the search and screening process, ideally in a flow diagram, and name studies that looked eligible but were excluded, with reasons.
17	Study characteristics	Cite each included study and summarize its design, participants, and setting.
18	Risk of bias in studies	Present the risk-of-bias judgment for every included study.
19	Results of individual studies	Give summary statistics and effect estimates, with confidence intervals, for each study and outcome.
20a-20d	Results of syntheses	Summarize the studies feeding each synthesis, present the synthesis results, report heterogeneity exploration, and report sensitivity analyses.
21	Reporting biases	Present the missing-results bias assessment for each synthesis.
22	Certainty of evidence	Give the certainty rating for each outcome assessed.
23a-23d	Discussion	Interpret the results against other evidence, discuss limitations of the evidence and of the review process, and state implications for practice, policy, and research.
24a-24c	Registration and protocol	State registration details or that the review was not registered, where the protocol can be found, and any amendments made.
25	Support	Name funding sources and describe the funder’s role, if any.
26	Competing interests	Disclose competing interests for every author.
27	Availability of data, code, and materials	State what is publicly available, data collection forms, extracted data, analytic code, and where to find it.

The PRISMA 2020 Flow Diagram Explained

Item 16a asks for the flow diagram, and it is the single most recognizable piece of a PRISMA-compliant paper, the box-and-arrow figure that shows how a search of several thousand records narrowed down to the handful of studies actually included. PRISMA 2020 reorganized it into three stages. Identification captures every record found, split into database and register searches versus other methods like hand-searching reference lists, plus records removed before screening even starts (duplicates, records marked ineligible by automation tools). Screening tracks records screened, records excluded, reports sought for retrieval, and reports that couldn’t be retrieved, kept as a separate box from reports excluded after full-text review. Included shows the final study and report counts, split out for the review itself versus any meta-analysis.

Two versions of the template exist, one for a new review starting from scratch and one for an update that also tracks studies carried over from a previous version of the review. Reviewers building the diagram by hand often reach for the flowchart tools in Editverse’s systematic review service, but the mechanics of constructing one, box by box, are covered in more depth in Editverse’s step-by-step guide to PRISMA flow chart steps, which walks through populating each box from raw screening numbers.

A good flow diagram earns its keep as a consistency check as much as an illustration. If the number of included studies stated in the results text doesn’t match the bottom box of the diagram, or a large batch of records disappears between screening and eligibility with no reason given, that mismatch is usually the first thing a statistical reviewer or editor notices, and it’s one of the more common reasons a paper gets sent back for revision before substantive peer review even starts.

A Worked Example: Writing Up Your Search Strategy (Item 7)

Item 7 is worth a closer look because it is where the gap between the 2009 and 2020 checklists is widest, and where authors most often under-report even when they intend to comply. Its actual requirement is short, and it asks for the full search strategy, not a paraphrase of it, for every database, register, and website searched, including any filters or date limits applied.^1,9 PRISMA-S, a companion extension focused specifically on search reporting, spells out sixteen elements a complete search description should include, among them the exact database name and platform, the date the search was run, and the full string of terms and Boolean operators used, for each source searched, not just the primary one.⁹

A search description that falls short of this usually reads along these lines. “We searched PubMed, Embase, and Cochrane CENTRAL for relevant studies on telehealth interventions in diabetes management, using terms related to telehealth and diabetes.” That sentence tells a reader what topic was searched but gives no way to reconstruct the actual search, evaluate whether it was too narrow, or update it later.

A version that satisfies item 7 reads quite differently. “MEDLINE was searched via Ovid from database inception to March 14, 2026, using the strategy (telehealth OR telemedicine OR ‘remote monitoring’).mp. AND (diabetes mellitus OR diabetes).mp., limited to English-language records. Equivalent strategies adapted for Embase (via Ovid) and Cochrane CENTRAL are provided in full in Supplementary File 1, along with the search dates for each database.” That version names the platform, states the date range, shows the actual syntax, and points to where the other two full strategies live, which is what lets a peer reviewer, per the PRESS guideline for evaluating electronic search strategies, actually check the search for missed synonyms or an overly narrow date filter rather than taking the topic description on faith.¹⁰

How PRISMA 2020 Connects to PROSPERO Registration

Items 24a through 24c ask authors to report where a review was registered, where its protocol can be found, and what changed between registration and publication. PROSPERO, the international register run by the University of York’s Centre for Reviews and Dissemination, is the destination most health-related reviews register with, and it has logged tens of thousands of protocols since launching in 2011.^7,8 Registering before screening begins creates a public record of the planned outcomes, comparisons, and analysis methods, which matters because it gives readers a way to check whether the published review’s outcomes match what was originally proposed, rather than outcomes that shifted after the results came in.

That relationship isn’t just procedural. An analysis of orthodontic systematic reviews found that PROSPERO-registered reviews scored measurably higher on the AMSTAR quality assessment tool than unregistered ones, even after accounting for other study characteristics.⁶ That doesn’t prove registration alone causes better reviews; teams organized enough to register a protocol in advance are probably also more methodologically careful in general. Either way, PRISMA 2020 treats registration status as something worth disclosing, since a reader benefits from knowing whether a plan existed before the results did. Editverse’s PROSPERO registration guide covers what the submission form actually asks for and how long registration typically takes.

Common PRISMA 2020 Compliance Gaps

Adoption of PRISMA 2020 has been slower than the guideline’s authors likely hoped. A 2025 audit of systematic reviews of interventions published after the checklist took effect found that fewer than a third of eligible reviews had switched to PRISMA 2020 at all, many still citing the 2009 version, and among those that had adopted it, average adherence measured against the full expanded checklist was 42.6%, with not a single review in the sample achieving full compliance.⁴ A separate audit of reviews in rehabilitation journals found that weaker adherence to PRISMA correlated with a higher risk of bias overall, suggesting the checklist gaps aren’t just cosmetic.⁵

A handful of items account for a disproportionate share of the gaps editors and reviewers flag.

Search strategy limited to one database. The 2009 habit of detailing only the primary database’s strategy persists, even though item 7 now requires all of them.
Flow diagram numbers that don’t reconcile with the text. A count of included studies in the results section that doesn’t match the diagram’s final box is one of the fastest ways to trigger a reviewer query.
No stated method for excluded full-text reports. Item 16b expects named studies and reasons for exclusion at the full-text stage, not just a total count.
Certainty assessment mentioned but not described. Naming GRADE without explaining which domains were downgraded and why leaves items 15 and 22 only partially satisfied.
Registration claimed without a working link or number. A statement that a review “was registered” carries little weight without the register name and registration number required under item 24a.

PRISMA vs CONSORT vs STROBE: Picking the Right Guideline

PRISMA is built for systematic reviews and meta-analyses specifically, and it is the wrong checklist for a primary study. Randomized trials follow CONSORT, which was substantially rewritten in 2025 and now runs to 30 items covering trial registration, participant flow, and harms reporting. Observational studies, cohort, case-control, and cross-sectional designs, follow STROBE instead, which swaps CONSORT’s randomization and blinding items for confounding and selection-bias considerations relevant to non-randomized designs. All three share the same underlying goal, enough methodological detail that a skeptical reader could evaluate, and in principle reproduce, what the authors actually did.

Reviews that pool data across multiple trials or observational studies sometimes need PRISMA layered alongside considerations from these study-level guidelines, since the quality of the underlying evidence still depends on how well each included study was designed and reported. Editverse’s guide to levels of evidence in clinical research covers how study design maps onto evidence hierarchies, which is a useful companion when deciding how much weight a given included study should carry in the synthesis.

How Editverse Supports PRISMA-Compliant Systematic Reviews

Passing a PRISMA 2020 check on the first submission usually comes down to whether the checklist was used as a drafting guide from the start rather than a box-ticking exercise the week before submission. Editverse’s systematic review team works with authors at both stages, from protocol development through to a completed manuscript.

Protocol development and PROSPERO registration support
Search strategy design and documentation across all databases, formatted to satisfy item 7
Building or auditing the flow diagram against the actual screening numbers
GRADE certainty assessments and Summary of Findings tables
A full pre-submission check against the current 27-item checklist, item by item

Working on a systematic review or meta-analysis?

Editverse checks manuscripts against the current PRISMA 2020 checklist and helps close specific gaps before a reviewer finds them.

Get PRISMA Compliance Support

Frequently Asked Questions About PRISMA 2020

How many items are in the PRISMA 2020 checklist?

Twenty-seven main items, several of which split into lettered sub-items (10a-b, 13a-f, 16a-b, 20a-d, 23a-d, and 24a-c), for 42 individual reporting points in total.

Is PRISMA 2020 mandatory for publication?

PRISMA itself has no enforcement power, but most journals following ICMJE recommendations require or strongly request a completed checklist alongside a systematic review submission, and editors commonly return a manuscript for PRISMA-specific revisions before it goes to peer review.

What’s the difference between PRISMA 2009 and PRISMA 2020?

PRISMA 2020 requires full search strategies for every database rather than just one, adds new items for protocol registration and certainty of evidence, splits the synthesis-methods and results-of-synthesis items into detailed sub-items, and restructures the flow diagram to separate records found through different methods.

Do I need to register my review before starting?

PRISMA 2020 doesn’t require registration, but items 24a-24c require disclosing whether the review was registered. Most reviewers register with PROSPERO before screening begins, since it is what items 24a-24c expect authors to report on either way.

Does PRISMA apply to scoping reviews or diagnostic accuracy reviews?

PRISMA’s core 27-item checklist is built for standard systematic reviews. Scoping reviews should use the PRISMA-ScR extension, and diagnostic test accuracy reviews should use PRISMA-DTA, both of which adapt specific items to those designs.

What is the most commonly missed item on the checklist?

Full search strategy reporting for every database (item 7) and reconciling the flow diagram with the numbers stated in the results text (item 16a) are the two gaps that come up most often in adherence audits and editorial checks alike.

References

Page MJ, McKenzie JE, Bossuyt PM, Boutron I, Hoffmann TC, Mulrow CD, et al. The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. BMJ. 2021;372:n71. PMID: 33782057
Page MJ, Moher D, Bossuyt PM, Boutron I, Hoffmann TC, Mulrow CD, et al. PRISMA 2020 explanation and elaboration: updated guidance and exemplars for reporting systematic reviews. BMJ. 2021;372:n160. PMID: 33781993
Page MJ, McKenzie JE, Bossuyt PM, Boutron I, Hoffmann TC, Mulrow CD, et al. Updating guidance for reporting systematic reviews: development of the PRISMA 2020 statement. J Clin Epidemiol. 2021;134:103-112. PMID: 33577987
Ivaldi D, Burgos M, Oltra G, Liquitay CE, Garegnani L. Adherence to PRISMA 2020 statement assessed through the expanded checklist in systematic reviews of interventions: A meta-epidemiological study. Cochrane Evid Synth Methods. 2025;2(5):e12074. PMID: 40476264
Innocenti T, Feller D, Giagio S, Salvioli S, Minnucci S, Brindisino F, et al. Adherence to the PRISMA statement and its association with risk of bias in systematic reviews published in rehabilitation journals: A meta-research study. Braz J Phys Ther. 2022;26(5):100450. PMID: 36270163
Sideri S, Papageorgiou SN, Eliades T. Registration in the international prospective register of systematic reviews (PROSPERO) of systematic review protocols was associated with increased review quality. J Clin Epidemiol. 2019;100:103-110. PMID: 29339215
Booth A, Clarke M, Dooley G, Ghersi D, Moher D, Petticrew M, Stewart L. The nuts and bolts of PROSPERO: an international prospective register of systematic reviews. Syst Rev. 2013;2:2. PMID: 22587842
Page MJ, Shamseer L, Tricco AC. Registration of systematic reviews in PROSPERO: 30,000 records and counting. Syst Rev. 2018;7:32. PMID: 29463298
Rethlefsen ML, Kirtley S, Waffenschmidt S, Ayala AP, Moher D, Page MJ, Koffel JB. PRISMA-S: an extension to the PRISMA Statement for Reporting Literature Searches in Systematic Reviews. Syst Rev. 2021;10:39. PMID: 33499930
McGowan J, Sampson M, Salzwedel DM, Cogo E, Foerster V, Lefebvre C. PRESS Peer Review of Electronic Search Strategies: 2015 Guideline Statement. J Clin Epidemiol. 2016;75:40-46. PMID: 27005575
Sterne JAC, Savović J, Page MJ, Elbers RG, Blencowe NS, Boutron I, et al. RoB 2: a revised tool for assessing risk of bias in randomised trials. BMJ. 2019;366:l4898. PMID: 31462531
Sterne JA, Hernán MA, Reeves BC, Savović J, Berkman ND, Viswanathan M, et al. ROBINS-I: a tool for assessing risk of bias in non-randomised studies of interventions. BMJ. 2016;355:i4919. PMID: 27733354
Guyatt GH, Oxman AD, Vist GE, Kunz R, Falck-Ytter Y, Alonso-Coello P, Schünemann HJ. GRADE: an emerging consensus on rating quality of evidence and strength of recommendations. BMJ. 2008;336(7650):924-926. PMID: 18436948

Related Resources

Editverse Research Team | Updated August 2026