PRISMA flow diagram numbers: how to count each box
In a PRISMA 2020 flow diagram, every box is a count and each stage must reconcile with the one above it:
- Records identified (databases and registers) minus records removed before screening (duplicates, marked ineligible by automation tools, removed for other reasons) = records screened.
- Records screened minus records excluded = reports sought for retrieval.
- Reports sought minus reports not retrieved = reports assessed for eligibility.
- Reports assessed minus reports excluded, with reasons = reports of included studies, which you then group into studies included in the review.
The key 2020 change to remember: records and reports are counted up to eligibility, but the final box counts studies, and one study can have several reports. PRISMA 2020 asks you to report both numbers.
What each box counts
| Box (PRISMA 2020 wording) | What to count |
|---|---|
| Records identified from databases / registers | Every record retrieved by each search, before deduplication. Report the number per database or register where feasible (the template allows this). |
| Records removed before screening | Three separate counts: duplicate records removed; records marked as ineligible by automation tools; records removed for other reasons. |
| Records screened | Records whose title and abstract were screened. |
| Records excluded | Records excluded at title and abstract stage. Reasons are not required here. |
| Reports sought for retrieval | Full texts you tried to obtain. |
| Reports not retrieved | Full texts you could not obtain. |
| Reports assessed for eligibility | Full texts assessed against your eligibility criteria. |
| Reports excluded (with reasons) | Each reason with its own count. The reason counts must sum to the total excluded. |
| Studies included in review / reports of included studies | Two numbers: the studies, and the reports that describe them. |
If you also searched by other methods (websites, organisations, citation searching), the PRISMA 2020 template has a second column with its own identified, sought, not retrieved, assessed, excluded boxes, which feed into the same final included box. Updated reviews have their own template versions that also count studies from the previous version.
Worked example (illustrative numbers)
Common mistakes
- Counting identified records after deduplication. The top box is the raw total; duplicates are shown being removed.
- Mixing reports and studies. If the final box says 21 studies but 22 + 3 reports flowed into it, explain the difference (linked reports).
- Reasons that do not sum. 30 + 20 + 14 must equal the 64 reports excluded. Use one primary reason per report if your log allows several.
- Silently dropping not-retrieved reports. Show them; readers want to know what you could not obtain.
- Folding other-methods records into the database column. PRISMA 2020 keeps them separate so readers can see what the search found by itself.
- Mismatch with the text. Numbers in the abstract, results and diagram must be identical.
- Using the 2009 four-box diagram when the journal expects PRISMA 2020.
Arithmetic checks to run before submission
- Sum of per-source counts = records identified.
- Identified minus all three before-screening removals = screened.
- Screened minus excluded = sought.
- Sought minus not retrieved = assessed.
- Sum of exclusion reasons = reports excluded.
- Assessed minus excluded = reports of included studies (per column).
- Studies included is less than or equal to reports included, and any difference is explained.
Our free PRISMA flow diagram generator runs these checks and draws the diagram; the PRISMA flow diagram guide covers the wider context.
Free download: PRISMA 2020 numbers worksheet
A fill-in worksheet with every PRISMA 2020 box (both columns), a reasons tally, and the seven arithmetic checks as tick boxes, so the numbers add up before you draw the diagram.
Download the PDF Two A4 pages. No email needed.
Do this in Notion
The free PDF is enough to do the job on paper. If you would rather keep it in one place that updates as you work, the Scoping & Systematic Review Manager is a Notion template that logs each search and every screening decision, tallies exclusion reasons, and builds the PRISMA counts from those records so the flow adds itself up.
It is a one-time purchase: Core $15, or Plus $24 with the AI pack. See exactly what is inside on the template page, or save by buying it with related templates in a bundle.
Sources
- Page MJ, McKenzie JE, Bossuyt PM, et al. The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. BMJ 2021;372:n71. doi.org/10.1136/bmj.n71
- Page MJ, Moher D, Bossuyt PM, et al. PRISMA 2020 explanation and elaboration. BMJ 2021;372:n160. doi.org/10.1136/bmj.n160
- PRISMA flow diagram templates (official site). www.prisma-statement.org/
Frequently asked questions
How do you count records in a PRISMA flow diagram?
Count every record retrieved from each database and register before deduplication, then show duplicates and other removals before screening. Records screened is what remains.
What is the difference between records, reports and studies in PRISMA 2020?
A record is a title and abstract indexed in a database; a report is a document such as a journal article or registry entry; a study is the investigation itself, which may be described in several reports.
Do I need to give reasons for exclusions at title and abstract screening?
No. PRISMA 2020 asks for reasons for reports excluded at full-text eligibility, not for records excluded at title and abstract screening.
Why do my PRISMA numbers not add up?
Usually because identified records were counted after deduplication, exclusion reasons do not sum to the total, or reports and studies were mixed. Run the seven arithmetic checks on this page.
Where do citation-searching records go in PRISMA 2020?
In the separate 'other methods' column, which has its own sought, retrieved and assessed boxes and feeds into the final included box.