To make a PRISMA 2020 flow diagram, choose the template for your review and search sources, enter counts from your search and screening logs, reconcile exclusions, and distinguish included reports from studies. Then generate and export the diagram. AnswerThis’s free PRISMA flow diagram generator handles new reviews without an account; updated reviews need a different template.
TL;DR
Choose a PRISMA 2020 template for a new or updated review and the sources searched.
Count records, reports and studies separately; keep a report-to-study mapping.
Report duplicates, non-retrieval and full-text exclusions at their respective stages.
AnswerThis generates unlimited new-review diagrams in your browser and exports PNG, SVG, editable PowerPoint and JSON.
A balanced diagram does not establish methodological quality or compliance with the full reporting checklist.
What is a PRISMA 2020 Flow Diagram?
A PRISMA 2020 flow diagram is a visual record of how evidence moved through a systematic review. It shows records identified, removals before screening, records screened, reports sought and assessed, exclusions with reasons, and the studies and reports included. Its purpose is transparent reporting, not certification of review quality.
Use these labels consistently:
Record: A database or website entry describing a report, commonly a title and abstract.
Report: A document supplying information about a study, such as an article, preprint or register entry.
Study: The investigation itself. One study can have multiple reports.
Labelled example: Identification: 1,000 records. Before screening: 150 duplicates removed. Screening: 850 records assessed and 700 excluded. Retrieval: 150 reports sought and 10 not retrieved. Eligibility: 140 reports assessed and 110 excluded with reasons. Inclusion: 30 reports describing 20 studies.
These illustrative counts adapt the generator’s published example, with the study total corrected to 20. They are not findings from a completed clinical review. The steps below show how they reconcile. Replace them with your own logs when reporting a real review.
How Does PRISMA 2020 Differ From PRISMA 2009?
The PRISMA 2020 statement replaces PRISMA 2009. The newer diagram separates stages and counting units more explicitly, with layouts for new and updated reviews.
Removals before screening
What to report in PRISMA 2020: Duplicate records, records marked ineligible by automation and other removals, where applicable
Retrieval
What to report in PRISMA 2020: Reports sought and reports not retrieved, separately from full-text eligibility decisions
Identification routes
What to report in PRISMA 2020: Databases/registers and, when applicable, other methods such as websites or citation searching
Final inclusion
What to report in PRISMA 2020: Studies included and reports of included studies as distinct counts
Review updates
What to report in PRISMA 2020: Previous-review evidence and evidence identified during the update, using the appropriate layout
Do not relabel an old diagram while leaving out stages your current logs contain. The diagram is one reporting item within the PRISMA checklist, not a substitute for reporting search strategies, appraisal or synthesis methods.
Which PRISMA 2020 Template Should You Use?
Choose on two dimensions: new or updated review, then databases/registers only or additional identification methods. The four official PRISMA 2020 templates are available as editable Word downloads.
First
Identification sources: Databases and registers only
Template: New-review, databases/registers template
Second
Identification sources: Databases/registers and other sources
Template: New-review template with other-sources arm
Third
Identification sources: Databases and registers only
Template: Updated-review, databases/registers template
Fourth
Identification sources: Databases/registers and other sources
Template: Updated-review template with other-sources arm
An updated review revisits an earlier review, not merely a draft with a revised search. Keep the prior evidence set identifiable when choosing an updated-review template. If you searched websites, organisations or reference lists, document that route rather than folding it silently into database totals.
How Do You Create a PRISMA 2020 Flow Diagram?
Before starting, gather dated search exports, a deduplication log, screening decisions, retrieval records, full-text exclusion reasons, and a report-to-study mapping. Resolve reviewer disagreements according to your protocol before counting the final decisions.
The worked example illustrates AnswerThis’s output, with the included-study count. It demonstrates counting only; no original review dataset was supplied.
1. Record identification counts by source
For an example, enter PubMed = 620 and Scopus = 380 in Records found per database. Those counts add to 1,000 identified records. Keep source-level totals before deduplication; do not use the number left after combining exports as the identification total.
In your own log, retain the database and interface, exact executed query, search date, filters and retrieved count. A draft query alone is not evidence of a completed search.
2. Account for removals before screening
Enter 150 in Duplicates removed. The example uses zero for Removed by automation and Removed for other reasons.
The calculation is 1,000 − 150 − 0 − 0 = 850 records screened. Duplicate records point to the same report. Separate publications about one trial are not automatically duplicate records; preserve them and map them to the study later.
3. Enter title-and-abstract screening decisions
Enter 850 in Records screened and 700 in Records excluded. That leaves 850 − 700 = 150 reports sought for retrieval in this example.
Use finalized decisions from the screening log. Do not subtract one reviewer’s exclusion count from another reviewer’s inclusion count or count unresolved conflicts twice.
4. Separate non-retrieval from eligibility exclusion
Of the 150 reports sought, 10 were not retrieved. Enter 10 in Reports not retrieved and 140 in Full-text reports assessed.
The calculation is 150 − 10 = 140 reports assessed for eligibility. A report you could not obtain is not a report whose contents you assessed and excluded. Retain the retrieval attempts in your review records.
5. Enter full-text exclusion reasons
The example uses one primary reason for each excluded report:
Wrong population
Reports excluded: 60
Wrong intervention
Reports excluded: 35
Wrong study design
Reports excluded: 15
Total
Reports excluded: 110
Enter each reason on its own line in Full-text exclusion reasons, for example, Wrong population = 60. The reasons total 60 + 35 + 15 = 110, leaving 140 − 110 = 30 included reports.
Your reasons should reflect your prespecified eligibility rules. Keep more detail in the screening log if a report fails several criteria, but avoid counting it under several primary reasons in the diagram.
6. Count studies independently from reports
Enter 20 in Studies included. Enter this count independently: one study may have several reports.
Enter 30 in Included reports. This is the count after full-text exclusions; count shared reports once. The illustrative example contains 30 reports describing 20 distinct studies. The tool cannot derive the study count from report arithmetic alone.
Keep a mapping with report ID, study ID, and the basis for linking them. Do not merge similar reports merely because their titles, authors, or interventions resemble one another.
7. Generate, inspect and export
For a new systematic review, choose PRISMA 2020 · systematic review under Review type and the appropriate Search sources option. Click Generate diagram, inspect the preview, then choose a format from Download.
PNG
Use it for: A flat image for a document or presentation
SVG
Use it for: Vector artwork that scales without raster pixelation
Editable PowerPoint
Use it for: Editing diagram text, boxes, and arrows
JSON
Use it for: Saving counts for import and later regeneration in the tool
View and download the labelled SVG example. The educational diagram adapts PRISMA 2020, credited to Page MJ et al., BMJ 2021;372:n71, under CC BY 4.0. Diagram description: 1,000 identified records become 850 screened records, 140 assessed reports, and 30 included reports describing 20 studies.
AnswerThis processes the diagram counts locally in your browser. Download the JSON before closing the tab if you want to preserve the counts for later regeneration. The diagram does not transfer automatically into an AnswerThis project, and the generator does not run searches, deduplicate references, or screen papers.
If your review uses both databases/registers and other sources, reconcile the counts within each identification route and check for shared reports before finalizing the included-report count. Count each shared report only once.
Do not infer the number of unique studies from the number of unique reports. Resolve study identity using your report-to-study mapping.
When Should You Use PRISMA-ScR for a Scoping Review?
Use PRISMA-ScR for reporting a scoping review. Its checklist has 20 essential items and 2 optional items. Scoping reviews map the scope and characteristics of evidence rather than necessarily estimating an intervention’s effect.
The generator’s scoping option adapts a PRISMA 2020 layout and uses scoping-review terminology, including sources of evidence included. It is not a separate official PRISMA 2020-ScR template. Confirm that the labels and counting units reflect your evidence types, and complete the applicable checklist separately.
What PRISMA 2020 Flow Diagram Mistakes Should You Avoid?
Counting reports as studies. Keep both totals and a report-to-study mapping.
Removing duplicates twice. Reconcile source exports with the deduplication log before screening counts are finalized.
Putting unavailable full texts under eligibility exclusions. Use the separate non-retrieval stage.
Counting one excluded report under several reasons. Assign one primary reason; retain details in the log.
Adding both identification arms without checking overlap. Count a shared included report once in the final total.
Leaving required counts blank. In this generator, blank is not zero. Enter zero only when the actual count is zero.
Using a new-review generator for an update. Use the official updated-review template instead.
Calling a diagram compliant because the arithmetic balances. Arithmetic cannot validate eligibility decisions, search completeness, or appraisal.
After checking the diagram, you can use the reporting checklist checker as a revision aid. It is free to try, not certification or a study-quality assessment. Review its findings against the official guideline and your complete manuscript; partial text cannot establish what the whole report contains.
FAQs
How do I make a PRISMA 2020 flow diagram?
Choose the template for your review type and sources, enter counts from finalized search and screening logs, reconcile exclusions, and distinguish reports from studies. Generate the diagram, inspect every label, and export it; keep editable counts for later corrections.
Is there a free PRISMA flow diagram generator?
Yes. AnswerThis PRISMA flow diagram generator is free, unlimited, and requires no account. It runs in the browser, supports new systematic reviews and a scoping adaptation, and exports PNG, SVG, editable PowerPoint, and JSON. Updated review templates are not currently supported.
Which PRISMA 2020 template should I use?
Select new or updated review, then databases/registers only or databases/registers plus other sources. Those two decisions determine which of the four official templates fits your review.
What changed in CONSORT 2025?
CONSORT 2025 provides a 30-item checklist for reports of randomized trials. Changes include a new open-science section and additional reporting on data sharing, patient/public involvement, harms, and handling missing data. CONSORT tracks trial participants; PRISMA tracks review evidence.
What is a systematic review?
A systematic review uses explicit methods to identify, select, appraise, and synthesize evidence addressing a defined question. A meta-analysis can be included but is not required. The PRISMA flow diagram reports selection, not the entire review method.
How do I build a Boolean search string for PubMed?
Group synonyms with OR, combine essential concepts with AND, and use parentheses to keep the groups together. Use NOT cautiously, inspect PubMed Search Details, and test retrieval of known relevant papers. Save the exact executed query, date, filters, and result count; a draft query is not a completed search.
How long does a systematic review take?
There is no single defensible completion time for every review. Scope, record volume, reviewer availability, full-text retrieval, and synthesis needs determine the schedule. Plan each stage separately; drawing the flow diagram is not the same as completing the review.
FAQ sources: PubMed User Guide and CONSORT 2025 statement.
For search preparation, the Search String Builder drafts queries for 12 search interfaces, including PubMed and Scopus presets. It does not run searches. Check the query in the destination interface and save the executed strategy, date, and count.
Next step: Reconcile your screening log, select the correct template, and export both a presentation format and an editable record of the counts. Then review the complete reporting checklist separately.
