PRISMA 2020 Explained: What Changed, Why It Matters, and How to Comply
PRISMA 2020 overhauled how systematic reviews are reported. This article breaks down every major change, walks you through the updated checklist and flow diagram, and shows you how to avoid the compliance mistakes that trigger desk rejections.
Dr. Sarah Mitchell
March 27, 2026
Validate your review against the full PRISMA 2020 checklist with our free tool at /resources/prisma-checklist-tool.
Key Takeaways
PRISMA 2020 replaces the original 2009 statement and applies to all systematic reviews, not just those with meta-analyses
The updated checklist contains 27 items with multiple sub-items that demand greater methodological transparency
The new flow diagram distinguishes between database and register sources versus other identification methods
Authors must now report certainty of evidence assessments such as GRADE for each primary outcome
Protocol deviations must be explicitly reported and justified in the final manuscript
The abstract checklist has expanded to 12 structured items to ensure key information is discoverable
Using PRISMA 2020 from the protocol stage prevents scrambling to retrofit reporting at manuscript submission
The PRISMA 2020 statement is now the definitive reporting standard for systematic reviews and meta-analyses worldwide. Published by Page et al. in The BMJ in 2021, it replaced the original PRISMA 2009 guidelines that had shaped evidence synthesis reporting for over a decade. If you are writing, reviewing, or editing a systematic review in 2026, understanding every change and its practical implications is not optional; it is essential for getting your work published.
This guide breaks down the complete PRISMA 2020 framework, explains the rationale behind each major update, walks through the redesigned flow diagram, and provides actionable guidance for achieving full compliance from protocol stage through manuscript submission.
The original PRISMA 2009 statement was a landmark achievement in evidence-based medicine. It standardised systematic review reporting, was endorsed by hundreds of journals, and became the most cited reporting guideline in research history. But a decade of widespread use revealed critical gaps that could no longer be patched with minor amendments.
Incomplete reporting persisted despite the guidelines. Multiple studies assessing PRISMA 2009 adherence across hundreds of published systematic reviews consistently found that authors failed to report key methodological details. Search strategies were summarised rather than fully presented. Risk-of-bias assessments were mentioned but not detailed. Certainty-of-evidence evaluations were omitted entirely from the majority of reviews. The original checklist was simply not granular enough to prevent these omissions.
Methodology advanced significantly between 2009 and 2020. New tools and frameworks emerged that the original statement never anticipated, including RoB 2 for randomised trial bias assessment, ROBINS-I for non-randomised studies, the GRADE framework for certainty of evidence, and the Synthesis Without Meta-analysis (SWiM) guideline for narrative synthesis. These advances required corresponding reporting guidance that PRISMA 2009 could not provide.
The landscape of evidence synthesis diversified. Network meta-analyses, scoping reviews, living systematic reviews, and individual participant data meta-analyses all became established methodologies with their own analytical requirements. The original PRISMA framework, designed primarily for conventional pairwise meta-analyses, needed a modular structure that could accommodate this growing family of review types through extensions.
The original flow diagram was oversimplified. It did not distinguish between records identified through database searching and those found through other methods like citation searching, grey literature, or expert contacts. This created an incomplete and sometimes misleading picture of the identification process, hiding important methodological decisions from readers and reviewers.
The Updated Checklist: What Changed and Why
PRISMA 2020 checklist sections; red tags mark items new or substantially revised from PRISMA 2009. Source: Page et al., 2021, BMJ.
The PRISMA 2020 checklist contains 27 items, many with multiple sub-items that demand substantially greater methodological detail than their 2009 counterparts. While the total item count matches the original, the content has been so extensively revised that treating PRISMA 2020 as a minor update would be a serious mistake. Below are the most significant changes, organised by the section of your manuscript they affect.
Title, Abstract, and Introduction (Items 1–3)
The title requirement remains straightforward: identify the report as a systematic review, meta-analysis, or both. The abstract, however, now must follow a structured 12-item format covering objectives, eligibility criteria, information sources, risk-of-bias assessment, synthesis methods, results including the number of studies and participants, certainty assessments, and interpretation. This expanded abstract ensures that readers can determine the scope and rigour of your review without accessing the full text. This is an increasingly important consideration as AI-powered search engines and abstract databases become primary discovery tools.
The introduction must describe the rationale in the context of existing knowledge (including other systematic reviews on the topic) and state exact objectives with explicit reference to the PICO elements or equivalent framework.
Methods Section (Items 4–16): The Heart of the Update
This is where the most consequential changes occur, and where most compliance failures happen.
Item 4: Protocol and registration is now a mandatory reporting element, not an optional afterthought. You must state whether a protocol was prepared, where it is accessible (registration number, DOI, or URL), and describe any amendments made between protocol registration and final reporting. If no protocol exists, you must explicitly state this. Silence on the topic is no longer acceptable.
Item 5: Eligibility criteria must specify inclusion and exclusion criteria with reference to PICO elements (or the equivalent framework for your review type), study characteristics such as design and minimum sample size, and report characteristics including language restrictions, publication status, and year of publication.
Items 6 and 7: Information sources and search strategy received major expansions. Item 6 requires you to list all information sources searched, including databases, registers, websites, organisations, reference lists, and other sources, along with the date of the last search for each. Item 7 requires the full search strategy for at least one database to be presented, ideally in a supplementary file with strategies for all databases. This single requirement addresses one of the most persistent reporting gaps under PRISMA 2009: vague search descriptions that made replication impossible.
Pro Tip: Start your PRISMA checklist at the protocol stage, not manuscript submission. Open the checklist when you write your protocol and map each item to a section of your methods. This approach ensures you collect the right information throughout the review (screening counts, exclusion reasons, risk-of-bias domain judgements, GRADE domain ratings) rather than scrambling to reconstruct it months later from incomplete notes. A checklist started at the protocol stage is always more accurate and complete than one filled in retroactively.
Items 8 and 9: Selection and data collection require detailed descriptions of the screening process (including the number of reviewers at each stage and how disagreements were resolved) and the data extraction process (including piloting of forms, number of extractors, and methods for obtaining or confirming data from study investigators).
Item 13: Synthesis methods is the most expanded section in the entire checklist and warrants careful attention. It requires you to describe:
The criteria for deciding which studies were eligible for each synthesis
How data were prepared for synthesis (e.g., converting means to medians, handling multi-arm trials)
Methods for tabulating and visually displaying results of individual studies and syntheses
Statistical methods for meta-analysis, including handling of heterogeneity
Methods for investigating possible causes of heterogeneity (subgroup analyses, meta-regression)
Sensitivity analyses planned and conducted
Methods for assessing reporting biases across studies
This granularity ensures that readers can evaluate not just what you found, but how you arrived at your findings, which is a fundamental principle of reproducible research.
Need help with your systematic review?
Our team builds comprehensive search strategies, handles screening across databases, and delivers PRISMA 2020-compliant systematic reviews.
PRISMA 2020 collapses four phases into three and tracks records and reports separately. Source: Page et al., 2021, BMJ.
The PRISMA 2020 flow diagram is one of the most visible changes and the element most frequently done incorrectly. Understanding the new structure is essential because editors will desk-reject manuscripts that use the outdated 2009 diagram.
Two Parallel Identification Tracks
The most significant structural change is the introduction of two parallel identification pathways at the top of the diagram. The left track captures records identified through databases and registers, meaning your traditional bibliographic database searching. The right track captures records identified through other methods, meaning citation searching, grey literature, expert contacts, websites, and any other non-database source.
Each track flows independently through its own identification and retrieval stages before converging at the eligibility assessment and inclusion stages. This dual-track structure makes transparent a practice that most rigorous reviews already follow but that the 2009 diagram obscured: comprehensive reviews identify studies through multiple channels, not just database searching.
New Boxes You Must Include
The 2020 flow diagram introduces several new reporting boxes that reflect modern systematic review practices:
Box
Purpose
Records removed before screening
Captures duplicates and records removed by automation tools (e.g., deduplication in Endnote, machine learning classifiers in ASReview)
Reports not retrieved
Documents studies where the full text could not be obtained, with a count and reason
Exclusion reasons for full-text reports
Requires categorisation of every full-text exclusion by reason (wrong population, wrong intervention, wrong outcome, wrong study design, wrong comparator)
Reports from other methods
Separately tracks records from citation searching, grey literature, expert contacts, and websites
Pro Tip: Use the official PRISMA 2020 flow diagram template. Download the editable template from the official PRISMA website rather than recreating it from scratch. The official version includes all required boxes in the correct layout, and editors will immediately flag custom diagrams that deviate from the standard structure. Templates are available in editable formats for Word, PowerPoint, and Lucidchart. Track your record counts in real time from the first day of screening. Reconstructing exact numbers months later is both tedious and error-prone.
Practical Guidelines for an Accurate Flow Diagram
Populate your flow diagram with exact numbers in every box, recorded contemporaneously throughout the review process. Use your screening software's export function to generate accurate counts at each stage. If your review updates a previously published systematic review, use the optional extended template that separately tracks previously identified studies carried forward into the update versus newly identified studies.
Reporting Certainty of Evidence with GRADE
One of the most consequential additions to PRISMA 2020 is the mandatory requirement to assess and report certainty in the body of evidence for each outcome. The GRADE (Grading of Recommendations, Assessment, Development and Evaluations) approach is the most widely used framework for this purpose and is required by Cochrane, WHO, and dozens of national guideline bodies.
How GRADE Works
GRADE evaluates five domains that can lower the certainty rating of a body of evidence. Starting from "high" certainty for randomised controlled trials or "low" certainty for observational studies, each domain can downgrade the rating by one or two levels:
Domain
What It Evaluates
Common Downgrade Triggers
Risk of bias
Methodological limitations across included studies
Majority of evidence from studies at high risk of bias in key domains
Inconsistency
Consistency of results across studies
Large, unexplained I-squared; conflicting direction of effects; non-overlapping confidence intervals
Indirectness
Directness of evidence to the review question
Population, intervention, comparator, or outcome differs from the review question
Imprecision
Precision of the pooled estimate
Wide confidence interval that crosses the threshold of clinical significance; few events or small total sample
Publication bias
Risk that studies with certain results are missing
Asymmetric funnel plot; significant Egger's test; known industry funding bias
In rare circumstances, three upgrading factors (large magnitude of effect, dose-response gradient, and all plausible confounders working against the observed effect) can increase certainty for observational evidence.
Summary of Findings Tables
Present your GRADE assessments in a Summary of Findings (SoF) table for each primary outcome. These tables display the pooled effect estimate, absolute and relative effects, the number of studies and participants contributing to each outcome, the GRADE certainty rating, and footnotes explaining each downgrade decision. Many journals now require SoF tables, and Cochrane reviews have mandated them for over a decade. GRADEpro GDT is the standard software tool for creating these tables.
Pro Tip: Report what you did, not just what you found. The biggest conceptual shift in PRISMA 2020 is the emphasis on transparency about process. For every methodological step (screening, extraction, risk-of-bias assessment, synthesis), describe your actual approach, any deviations from your protocol, and the reasons for those deviations. Reviewers and editors now expect this level of procedural transparency, and manuscripts that only report results without adequate methodological detail will be returned for major revision or rejected outright.
Protocol Amendments and Deviations
PRISMA 2020 creates a structured audit trail from protocol to publication through three interconnected reporting requirements. First, you must state whether a protocol was registered or published and provide access details (Item 4). Second, you must describe and justify any deviations from the registered protocol (Item 13b). Third, you must report any analyses that were not pre-specified, clearly distinguishing them from planned analyses (Item 13d).
This framework does not punish amendments. Protocols are living documents, and legitimate methodological changes are expected as reviewers learn more about the literature landscape. The standard is transparency, not rigidity. The critical distinction is between changes driven by methodological learning (acceptable when documented) and changes driven by results (which undermine the credibility of the entire review).
Need a PRISMA 2020 compliant systematic review? Our methodologists ensure every item is addressed. Get a free quote at /get-a-quote.
Common Compliance Failures and How to Avoid Them
Published assessments of PRISMA 2020 adherence across hundreds of systematic reviews have identified consistent patterns of non-compliance. Understanding these common failures allows you to avoid them proactively rather than addressing them during peer review revisions.
Compliance Failure
How to Avoid It
Missing full search strategy
Provide the complete strategy for at least one database in a supplementary file, with all terms, operators, limits, and date of execution
Using the 2009 flow diagram
Download and use the official 2020 template; include both identification tracks and all new required boxes
No certainty assessment
Conduct GRADE (or equivalent) for every primary outcome and present results in Summary of Findings tables
Vague risk-of-bias reporting
Name the specific tool, report domain-level results for each study (traffic light plots), and explain how assessments informed your synthesis
Missing protocol information
Include the registration number, registry name, and access URL; if no protocol exists, state this explicitly
No deviation reporting
Compare your final methods against your protocol point by point and report all differences with justifications
Oversimplified limitations
Discuss limitations of the evidence, the review process, and the conclusions as three distinct topics
Pro Tip: Prepare the PRISMA checklist as a supplementary file during manuscript writing. Many journals require the completed checklist with page or section numbers at submission. Build this document incrementally as you draft each section of your manuscript, filling in the checklist item and page number immediately after writing each corresponding section. Adding all page references retroactively after the manuscript is complete is tedious, error-prone, and almost always results in missing or incorrect references.
PRISMA 2020 Extensions for Specialised Reviews
PRISMA 2020 is the core statement, but a growing family of extensions addresses the unique reporting needs of specific review types. Using the appropriate extension alongside the core statement ensures your specialised review meets the full reporting standard expected by journals and peer reviewers.
Extension
Full Name
Applies To
PRISMA-S
PRISMA for Search Strategies
Reporting of search methodology across all review types
PRISMA-ScR
PRISMA for Scoping Reviews
Scoping reviews following JBI or Arksey-O'Malley framework
PRISMA-NMA
PRISMA for Network Meta-Analyses
Reviews comparing three or more interventions simultaneously
PRISMA-DTA
PRISMA for Diagnostic Test Accuracy
Reviews evaluating sensitivity, specificity, and predictive values
PRISMA-IPD
PRISMA for Individual Participant Data
Reviews using individual-level data from included studies
PRISMA-P
PRISMA for Protocols
Protocol reporting (compatible with PRISMA 2020)
Integrating PRISMA 2020 Into Your Workflow
The most effective and least painful approach to PRISMA 2020 compliance is integrating it into your workflow from the first day of your review, not treating it as a manuscript-submission checkbox exercise. This front-loaded strategy transforms PRISMA from a burdensome compliance requirement into a quality-improvement tool that strengthens every stage of your review.
At the protocol stage, open the PRISMA 2020 checklist and map each of items 4 through 16 to a corresponding section of your protocol. Every methodological decision you pre-specify in your protocol directly satisfies a PRISMA reporting requirement, making the two documents mutually reinforcing rather than redundant.
During data collection, track your screening counts, exclusion reasons at every stage, and risk-of-bias domain judgements in real time using your review management software. Export these data regularly so that populating your flow diagram and risk-of-bias figures at the manuscript stage requires retrieval, not reconstruction.
During synthesis, prepare your GRADE Summary of Findings tables as you complete each meta-analysis or narrative synthesis, rather than deferring certainty assessments until after the manuscript draft is written. Assessing certainty during synthesis often reveals analytical gaps (missing sensitivity analyses, unexplored subgroups, or insufficient investigation of heterogeneity) that are easier to address before the manuscript is drafted.
At manuscript submission, verify completeness by walking through the checklist one final time and adding page or section numbers for every item. Submit the completed checklist as a supplementary file alongside your manuscript, your PRISMA 2020 flow diagram, your risk-of-bias summary figures, and your Summary of Findings tables.
This integrated approach reduces the total time spent on PRISMA compliance by eliminating the retroactive scramble that characterises most teams' first encounter with the updated guidelines, and it produces a more complete, more accurate, and more defensible systematic review.
PRISMA 2020 is the updated international reporting guideline for systematic reviews and meta-analyses, published by Page et al. in the BMJ in 2021. It replaces the original 2009 statement. The key differences include expanded checklist items with sub-items demanding greater detail, a redesigned flow diagram that distinguishes between database and non-database sources, mandatory reporting of certainty assessments, explicit requirements for protocol registration and deviation reporting, and an updated 12-item abstract checklist. The 2020 version reflects a decade of methodological advances and addresses gaps identified by researchers who found widespread incomplete reporting under the original guidelines.
No. Despite the name including meta-analysis, PRISMA 2020 applies to all systematic reviews regardless of whether they include a quantitative synthesis. The guidelines are equally relevant for systematic reviews with narrative synthesis, reviews of qualitative evidence, scoping reviews (with the PRISMA-ScR extension), and other types of evidence synthesis. The core principle is transparent, complete reporting of your methods and findings.
The most frequent compliance failures include omitting protocol registration details, providing incomplete search strategy documentation (failing to report full strategies for all databases), not distinguishing between database and other identification sources in the flow diagram, failing to report certainty of evidence assessments, not listing all excluded full-text studies with reasons, omitting a description of the methods used for risk-of-bias assessment, and neglecting to report protocol amendments and their justifications. These gaps often trigger desk rejections at journals that enforce PRISMA compliance.
Download the official editable template from prisma-statement.org. The 2020 flow diagram has two parallel tracks at the top: one for records identified through databases and registers, and another for records identified through other methods such as citation searching, grey literature, and expert contacts. Below these tracks, the diagram converges into screening, eligibility assessment, and final inclusion boxes. Each box should include the exact number of records at that stage, and exclusion boxes must list reasons. You can create the diagram in any software that supports editable shapes, including PowerPoint, Word, Lucidchart, or dedicated tools like the PRISMA flow diagram generator.
Yes, and enforcement is increasing. Many high-impact journals including The BMJ, The Lancet, JAMA, and Cochrane Database of Systematic Reviews require a completed PRISMA 2020 checklist at submission and will desk-reject manuscripts that show major non-compliance. Even journals that do not strictly enforce PRISMA 2020 at submission will have peer reviewers who use it as a benchmark. Incomplete reporting is one of the most common reasons for revision requests and rejections of systematic review manuscripts. Proactive compliance saves time and increases your chances of acceptance.
GRADE, or Grading of Recommendations, Assessment, Development and Evaluations, is a systematic framework for rating the certainty of a body of evidence. It evaluates five domains: risk of bias, inconsistency, indirectness, imprecision, and publication bias. Each domain can downgrade certainty from high to moderate, low, or very low. PRISMA 2020 requires reporting certainty assessments because knowing the pooled effect size is insufficient without understanding how confident we should be in that estimate. GRADE is the most widely used approach, but alternatives like CERQual for qualitative evidence are also acceptable.
PRISMA 2020 was designed for systematic reviews, but the PRISMA-ScR (PRISMA extension for Scoping Reviews) provides a parallel framework specifically for scoping reviews. Many items overlap, and using PRISMA 2020 principles for any evidence synthesis will improve your reporting quality. If you are conducting a scoping review, use PRISMA-ScR as your primary checklist and reference PRISMA 2020 for additional guidance on items that both frameworks share, such as search strategy reporting and flow diagram construction.
Share
Found this useful? Share it with your colleagues.
Need help with your systematic review?
Our team builds comprehensive search strategies, handles screening across databases, and delivers PRISMA 2020-compliant systematic reviews.
Building Your Search Strategy? We Can Run the Whole Review.
From protocol to PRISMA flow diagram, our PhD team handles database searches, screening, data extraction, and a publication-ready manuscript. Every step, done right.
Our promise: Free rework on search, screening, or synthesis if reviewers push back.
4.9 / 5Quote within a few hoursPRISMA 2020 + Cochrane HandbookPhD methodologistConfidential by default
Dr. Sarah Mitchell holds a PhD in Biostatistics from Johns Hopkins Bloomberg School of Public Health and has over 15 years of experience in systematic review methodology and meta-analysis. She has authored or co-authored 40+ peer-reviewed publications in journals including the Journal of Clinical Epidemiology, BMC Medical Research Methodology, and Research Synthesis Methods. A former Cochrane Review Group statistician and current editorial board member of Systematic Reviews, Dr. Mitchell has supervised 200+ evidence synthesis projects across clinical medicine, public health, and social sciences.
Research Gold delivers systematic reviews that meet every PRISMA 2020 and GRADE reporting requirement. Start your project at /get-a-quote.
Building Your Search Strategy? We Can Run the Whole Review.
From protocol to PRISMA flow diagram, our PhD team handles database searches, screening, data extraction, and a publication-ready manuscript. Every step, done right.
Meta-analysis in psychology pools the effect sizes from many studies into one reliable result. Learn the definition, real examples, and how researchers run one.
Human-written, AI-assisted, AI-screened: the labels have stopped being descriptive. Here is the single threshold journals actually use, what you must disclose, and where Research Gold draws the line.
