Skip to content

Research reporting

PRISMA flow diagram maker

Make the record-selection process visible with flowss's PRISMA engine. Enter the counts and exclusion reasons as source, then inspect the flow and the engine's checks. The counts in the example are teaching data; replace them with your review's own evidence.

From starting point to finished diagram

  1. Choose the flow that matches your review

    Start with the reference template. Use the existing PRISMA template collection for an updated review or a separate other-methods arm when that reflects your actual search.

  2. Replace every illustrative count

    Use the search and screening records from your review. State full-text exclusion reasons and counts, and retain the distinction between reports retrieved and studies included.

  3. Resolve the checks and review the figure

    Read each arithmetic error, warning or missing-term finding. Correct the underlying count or explicitly supply the missing term; then compare the exported figure with your review log.

PRISMA 2020 flow · Editable template

Example: a review with 116 reports and 98 studies

This reference example identifies 3,006 records, removes 814 before screening, screens 2,192, seeks 452 reports and assesses 421. Its 305 named full-text exclusions leave 116 reports describing 98 studies.

  • Check the screening arithmetic: 2,192 screened minus 1,740 excluded equals 452 reports sought. Of those, 31 are unavailable and 421 are assessed.
  • Keep reports and studies separate: several reports can describe the same study. The template's 116 and 98 are different quantities, not a discrepancy to erase.
Remote monitoring in heart failure — systematic review3,006 records identified · 3 exclusion reasonsIdentificationScreeningEligibilityIncludedRecords identified from:n = 3,006Databases: 2,840 · Registers: 120 · Other methods: 46Records removed before screeningn = 814Duplicates removed: 610Excluded by automation: 180Removed for other reasons: 24Records screened (title / abstract)n = 2,192Records excluded by reviewern = 1,740Reports sought for retrievaln = 452Reports not retrievedn = 31Reports assessed for eligibilityn = 421Reports excluded:n = 305No remote-monitoring arm: n = 148Wrong outcome measure: n = 96Conference abstract only: n = 61Studies included in reviewn = 98Reports of included studies: 116Every record is accounted for: all 4 stages of the flow balance.INFO116 reports describe 98 studies: 18 reports are a further paper about a study already counted.OKAll 13 stated counts are whole numbers of records.OKIdentification balances: 2,840 + 120 + 46 − 610 − 180 − 24 = 2,192 `title_screened`.OKScreening balances: 2,192 − 1,740 = 452 `abstracts_sought`.OKRetrieval balances: 452 − 31 = 421 `fulltext_assessed` (`other_sought` and `other_unavailable` are not stated and read as zero).OKEligibility balances: 421 − 305 = 116 `reports_included`.OKThe funnel descends: all 4 stage-to-stage comparisons hold — no count is larger than the one before it.OKAll 3 exclusion reasons are named and distinct, accounting for 305 reports.
Systematic Review — the Reference Flow. View the template and related examples.
Read and copy the editable source
title "Remote monitoring in heart failure — systematic review"

# Identification — where the records came from.
db_records 2840
register_records 120
other_records 46

# Identification — what left before anyone read a title.
duplicates 610
excluded_auto 180
excluded_other 24

# Screening.
title_screened 2192
title_excluded 1740

# Retrieval.
abstracts_sought 452
abstracts_unavailable 31

# Eligibility.
fulltext_assessed 421
exclude_reason "No remote-monitoring arm" 148
exclude_reason "Wrong outcome measure" 96
exclude_reason "Conference abstract only" 61

# Included. 116 reports describe 98 distinct studies.
reports_included 116
studies_included 98

A balanced diagram is one part of reporting

The engine can check the counts you supply and highlight missing terms. It cannot establish that your search was comprehensive, that eligibility decisions were correct, or that a manuscript meets every reporting requirement. Keep the search strategy and screening evidence with the review.

Do not force every review into one shape

An updated review can carry forward earlier reports, and an other-methods search can enter through a separate arm. The PRISMA engine reference explains those structures and the checks that apply to each.

Before you start

Access and limits

  • Opening the Studio requires an account. Free includes the Studio and keeps up to 3 saved figures in it and 3 cloud Library projects.
  • The PRISMA engine renders and checks its source without AI. Reading and copying the public template source is free; the Studio requires sign-in.
  • These checks assist reporting; they do not certify PRISMA compliance or validate a study's findings.
Compare plans