How to make a CONSORT flow diagram
- Choose the number of arms (1 to 4) and name them; the names follow “Allocated to”.
- Enter enrolment: participants assessed for eligibility and the reasons for exclusion. Randomised is calculated from the arms if you leave it empty.
- For each arm, enter allocation, follow-up and analysis, with a reason and number for everyone who did not receive the intervention, discontinued, was lost to follow-up or was excluded from the analysis.
- Fix any warnings above the diagram, such as arms that do not add up to the number randomised.
- Apply your journal’s rules and export a vector PDF or SVG, or PNG/TIFF up to 1200 dpi. “Results text” drafts the participant-flow paragraph for your Results section.
What reviewers check
- The numbers add up at every stage, in every arm, from assessment to analysis.
- Every loss has a reason, and losses are balanced (or explained) between arms.
- The analysed numbers match the analysis population stated in the methods (intention to treat or per protocol).
- The diagram and the text use the same numbers. For systematic reviews, use the PRISMA flow diagram instead; for the trials included in a review, see the risk-of-bias plot.
Frequently asked questions
What changed in the CONSORT 2025 flow diagram?
CONSORT 2025 (Hopewell et al., BMJ 2025) replaced CONSORT 2010. In the flow diagram, loss to follow-up and the analysed numbers now refer explicitly to the primary outcome (“Lost to follow-up for primary outcome”, “Analysed for primary outcome”), and discontinuation of the intervention has its own box, separate from loss of outcome data. CleanFigure uses this wording by default; every box can be reworded if your journal asks for it.
How are the missing numbers calculated?
Leave a box empty and it is filled in from the others: excluded is the sum of the reasons, randomised is the sum of the allocated participants (or assessed minus excluded), received is allocated minus those who did not receive the intervention, and analysed is allocated minus those lost to follow-up and excluded from the analysis. A number you type always wins, and a mismatch is shown above the diagram.
Does “discontinued intervention” reduce the number analysed?
No. Participants who stop the intervention usually still provide outcome data and, under the intention-to-treat principle, are analysed in their allocated group. Only participants without primary outcome data (lost to follow-up) or excluded from the analysis reduce the analysed number. If your analysis differs, type the analysed number and explain it in the excluded-from-analysis reasons.
Can I draw trials with more than two arms?
Yes, from one to four parallel arms. CONSORT says the diagram has no prescribed format beyond showing the numbers in each group at each stage, so each arm gets its own column with the same boxes. Cluster, factorial and crossover designs have CONSORT extensions with their own diagrams and are not covered yet.
Do I need to give reasons?
Yes. CONSORT asks for reasons for exclusions before randomisation, for not receiving the allocated intervention, for discontinuation, for loss to follow-up and for exclusion from the analysis, with a number for each. CleanFigure reminds you when a count has no reasons.
How do I cite the diagram?
Cite the CONSORT 2025 statement (Hopewell S, Chan AW, Collins GS, et al. BMJ 2025;389:e081123) for the template, and optionally CleanFigure. You can print the citation under the diagram.