CONSORT Flow Diagram for Randomised Trials

<role vocab="credit"
      vocab-identifier="https://casrai.org/dictionary/"
      vocab-term="CONSORT Flow Diagram for Randomised Trials"
      vocab-term-identifier="https://casrai.org/dictionary/term/consort-flow-diagram" />

{
  "@context": "https://schema.org",
  "@type": "DefinedTerm",
  "@id": "https://casrai.org/dictionary/term/consort-flow-diagram",
  "name": "CONSORT Flow Diagram for Randomised Trials",
  "identifier": "https://casrai.org/dictionary/term/consort-flow-diagram",
  "description": "The CONSORT flow diagram is the figure that accounts for every participant in a randomised trial across four stages — enrolment, intervention allocation, follow-up and data analysis — giving a count in each box and a reason for every departure. It is required by the CONSORT 2025 statement under item 22a, with losses and exclusions after randomisation under item 22b. The thing that decides whether a reviewer accepts your diagram is not its shape. It is whether the numbers reconcile and whether the drop-offs carry reasons. Two boxes account for most of the trouble: Did not receive allocated intervention and Excluded from analysis. Leaving either blank, or filling it with a bare number and no reason, is the standard way a diagram fails review — and it is exactly where a reader looks for evidence of bias. What CONSORT 2025 actually asks for CONSORT 2025, published 14 April 2025 simultaneously in The BMJ, JAMA, The Lancet, Nature Medicine and PLOS Medicine, is a 30-item checklist. The participant-flow items are: Item 22a — “For each group, the numbers of participants who were randomly assigned, received intended intervention, and were analysed for the primary outcome”. Item 22b — “For each group, losses and exclusions after randomisation, together with reasons”. Note what item 22a does not mandate: the number screened. The explanation and elaboration paper is explicit that eligibility-screening counts should be reported “if available”, that they bear on external validity only, and that they “may not always be known”. If you ran an opportunistic or registry-based recruitment where nobody counted the people who were looked at and passed over, say so rather than reverse-engineering a screening figure. CONSORT also declines to prescribe a shape. The E&E states that “although CONSORT strongly recommends using a flow diagram to communicate participant flow throughout the study, there is no specific, prescribed format”. The template is a starting point, not a form to be filled in unchanged. For the wider checklist context, see the full CONSORT checklist for RCT reporting. The four stages and the exact box labels The official CONSORT 2025 flow diagram template (a Word file distributed by the CONSORT/SPIRIT group) uses these labels for a two-group parallel trial: Enrolment Assessed for eligibility (n= ) Excluded (n= ), broken out as: Not meeting inclusion criteria (n= ); Declined to participate (n= ); Other reasons (n= ) Randomised (n= ) Allocation One column per arm, each containing: Allocated to intervention (n= ) Received allocated intervention (n= ) Did not receive allocated intervention (give reasons) (n= ) Follow-Up Lost to follow-up for primary outcome (give reasons) (n= ) Discontinued intervention (give reasons) (n= ) Analysis Analysed for primary outcome (n= ) Excluded from analysis (give reasons) (n= ) Four of these boxes say “give reasons”. That instruction is the substance of item 22b, and it is the part most often dropped when a diagram is redrawn by hand or rebuilt in a figure-editing tool. What changed from CONSORT 2010 The diagram is recognisably the same, but three differences matter when you are updating an old template: Item numbers moved. Participant flow was items 13a and 13b in CONSORT 2010; in CONSORT 2025 it is 22a and 22b. Journal submission checklists that still ask for “item 13” are on the old numbering. The follow-up and analysis boxes are now tied to the primary outcome. The CONSORT 2010 template read “Lost to follow-up (give reasons)” and “Analysed”. The 2025 template reads “Lost to follow-up for primary outcome (give reasons)” and “Analysed for primary outcome”. This removes a long-standing ambiguity about which outcome the denominators referred to. Wording aligned with SPIRIT. Item 22a says “received intended intervention” where CONSORT 2010 item 13a said “received intended treatment”, part of a general alignment of CONSORT and SPIRIT 2025 item wording. The counts the diagram has to carry The CONSORT 2025 E&E sets out, stage by stage, which counts are needed and why. In its own framing: Enrolment — people evaluated for potential enrolment, and those who did not meet the inclusion criteria or met them but declined. These “indicate whether trial participants were likely to be representative of all patients seen”; they are “relevant to assessment of external validity only, and they are often not available”. Randomisation — participants randomly assigned. Described as the “crucial count for defining trial size and assessing whether a trial has been analysed by intention-to-treat”. Treatment allocation — participants who did and did not receive intervention as allocated, by group, with reasons. “Important counts for assessment of internal validity and interpretation of results”. Follow-up — participants who completed, and did not complete, intervention as allocated and follow-up as planned, by group, with reasons. Analysis — participants included in and excluded from the main analysis of the primary outcome, by group, with reasons. Again the “crucial count for assessing whether a trial has been analysed by intention-to-treat”. A practical consequence: the diagram is a two-way reconciliation. Randomised should equal the sum of the arms’ allocated counts. Within each arm, allocated minus every subtracted box should equal analysed — and if it does not, the gap must be named somewhere, not absorbed silently. Loss to follow-up is not the same as discontinuing the intervention These are separate boxes because they are separate events, and conflating them destroys the reader’s ability to judge the result. A participant can stop taking the drug and still attend every visit, so their primary outcome is measured and they remain in the intention-to-treat analysis. Another can adhere perfectly and then move away, leaving no outcome data at all. The E&E is direct about this: attrition from loss to follow-up, “which is often unavoidable, needs to be distinguished from investigator-determined exclusion for such reasons as ineligibility, withdrawal from treatment, and poor adherence to the trial protocol”. It adds that “imbalances in such omissions between groups may be especially indicative of bias” — which is why the counts are demanded per arm rather than as a single trial total. That asymmetry is the mechanism behind attrition bias. Post-randomisation exclusions: the trap that breaks intention-to-treat The most consequential rule on the page concerns people you discover, after randomisation, should never have been randomised. The E&E for item 22b: “If participants were excluded after randomisation (contrary to the intention-to-treat principle) because they were found not to meet eligibility criteria, they should be included in the flow diagram. Use of the term ‘protocol deviation’ in published articles is not sufficient to justify exclusion of participants after randomisation. The nature of the protocol deviation and the exact reason for excluding participants after randomisation should always be reported.” In other words: an ineligible participant who was randomised still enters the diagram at “Randomised”, still appears in their allocated arm, and if they are dropped, appears again in “Excluded from analysis” with the specific reason. Quietly removing them from the top of the diagram makes the trial look cleaner than it was and is a departure from intention-to-treat analysis that a careful reader cannot detect. CONSORT 2025 pairs this with item 21b, which requires a stated “definition of who is included in each analysis (eg, all randomised participants), and in which group”. The E&E defines intention-to-treat as meeting two conditions — all randomised participants are included in the analysis, and each is retained in the group to which they were allocated. If your diagram’s analysed count is smaller than your randomised count, item 21b is where you owe the reader an explanation, and “modified intention-to-treat” is a label, not an explanation. The diagram does not settle the denominators for every outcome A frequent misreading is that the “Analysed” box fixes the N for the whole results section. It does not. The E&E for item 26 is explicit: “although the flow diagram (item 22a) should indicate the numbers of participants included in the analysis of the primary outcome, the number of participants with available data will often vary for different outcomes and at different time points”. CONSORT 2025 item 26 accordingly asks, for each primary and secondary outcome and by group, for both “the number of participants included in the analysis” and “the number of participants with available data at the outcome time point”. Those are two different numbers whenever imputation or a model-based analysis is used, and the diagram only carries the primary-outcome pair. Secondary-outcome denominators belong in the results tables. Adapting the diagram to your design The template assumes a two-group parallel trial. The E&E anticipates changes: “the exact form and content of the flow diagram may be varied according to specific features of a trial. For example, many trials of surgery or vaccination do not include the possibility of discontinuation.” For non-pharmacological interventions it notes it “may be important to report the number of care providers or centres performing the intervention in each group and the number of participants treated by each care provider or in each centre”. Two adaptations come with published extensions rather than being left to judgement: Cluster randomised trials. The CONSORT 2010 cluster extension adds a parallel set of counts at the cluster level: item 13a additionally requires, for each group, “the numbers of clusters that were randomly assigned, received intended treatment, and were analysed for the primary outcome”, and item 13b requires losses and exclusions “for both clusters and individual cluster members”. A cluster trial diagram therefore carries two counts in most boxes. See cluster randomised trials and the ICC for the design consequences. Other extensions. The CONSORT/SPIRIT group publishes separate flow diagram templates for pilot and feasibility trials, non-pharmacological treatments, within-person trials and CONSORT-Equity, alongside the main CONSORT 2025 diagram. Why journals push the diagram so hard The evidence behind the requirement predates CONSORT 2025 by two decades. Egger, Jüni and Bartlett examined 270 reports of randomised trials published in 1998 in five general and internal medicine journals. Only 139 (51.5%) included a flow diagram, with uptake ranging from 8.4% in the New England Journal of Medicine to 93.8% in The Lancet. Among the diagrams that did appear, only 73 (52.5%) gave the number of participants who received allocated interventions and only 32 (23.0%) gave the number included in the analysis. Overall completeness of reporting about participant flow was associated with publication of a flow diagram, and the authors proposed the revised structure that CONSORT adopted. The 2025 E&E repeats the finding: reporting of what happened to participants and their data was “considerably more thorough in articles that included a diagram of the flow of participants through a trial than in those that did not”. A checklist before you submit Randomised equals the sum of the allocated counts across arms. Within each arm, allocated minus every subtraction equals analysed — no unexplained residual. Every “give reasons” box has reasons, itemised, not a total. Losses to follow-up and intervention discontinuations are in separate boxes, per arm. Anyone excluded post-randomisation for retrospective ineligibility still appears in the diagram, with the specific reason — not “protocol deviation”. The analysed count matches the primary-outcome N reported in the results and in the abstract. The analysis population named in item 21b is consistent with the diagram’s arithmetic. Cluster trials show both cluster-level and participant-level counts. Frequently asked questions Is the CONSORT flow diagram mandatory? It is part of the CONSORT 2025 checklist under item 22a, and many journals require both the completed checklist and the diagram at submission. CONSORT’s own language is that it “strongly recommends” using a flow diagram; in practice, editors at journals endorsing CONSORT treat it as expected for a parallel-group randomised trial. The checklist item itself — the counts — is not optional even where the figure format is. Where do I get the official template? The CONSORT/SPIRIT group distributes the CONSORT 2025 flow diagram as an editable Word document from consort-spirit.org, alongside the editable 30-item checklist and extension-specific diagrams. Using the official file is worth it precisely because it preserves the “give reasons” prompts that hand-drawn versions tend to lose. Do I have to report the number screened for eligibility? Report it if you have it. CONSORT’s explanation says the number assessed for eligibility “should also be reported” if available, but is “relevant to external validity only and is arguably less important than the other counts”, and acknowledges it “may not always be known”. Do not invent it. If your recruitment route makes a screening denominator meaningless, state that instead. What is the difference between item 22a and item 22b? Item 22a is the positive accounting — how many were randomised, received the intervention, and were analysed for the primary outcome, per group. Item 22b is the negative accounting — the losses and exclusions after randomisation, per group, with reasons. In practice both are usually satisfied by the same figure, since the template has boxes for each. Which item number is participant flow in CONSORT 2025? Items 22a and 22b, under the heading “Participant flow, including flow diagram”. In CONSORT 2010 the same content sat at items 13a and 13b. If a journal’s submission system still refers to item 13, it is running the 2010 numbering. Can the analysed count differ from the randomised count? Yes, and it often does — but every participant in the difference has to be visible in the diagram with a reason. A strict intention-to-treat analysis includes all randomised participants in their allocated groups; the E&E acknowledges that strict adherence “is often difficult to achieve” because of missing outcomes or non-adherence, and requires that whichever analysis population you use is described under item 21b. How do I handle a three-arm or factorial trial? Add columns. The template is drawn for two groups but CONSORT states there is no prescribed format, and the counts required per group are the same regardless of how many groups there are. For factorial designs, be explicit in the caption about whether the columns represent randomised cells or marginal comparisons, since the analysed denominators differ between the two readings. References Hopewell S, Chan A-W, Collins GS, et al. CONSORT 2025 statement: updated guideline for reporting randomised trials. BMJ 2025;389:e081123. Hopewell S, Chan A-W, Collins GS, et al. CONSORT 2025 explanation and elaboration: updated guideline for reporting randomised trials. BMJ 2025;389:e081124. (Items 21b, 22a, 22b, 26 and table 7.) CONSORT/SPIRIT. CONSORT 2025 flow diagram and editable checklist downloads. consort-spirit.org. Egger M, Jüni P, Bartlett C, for the CONSORT Group. Value of flow diagrams in reports of randomized controlled trials. JAMA 2001;285:1996–9. Campbell MK, Piaggio G, Elbourne DR, Altman DG, for the CONSORT Group. CONSORT 2010 statement: extension to cluster randomised trials. BMJ 2012;345:e5661. EQUATOR Network. CONSORT reporting guideline record and extensions.",
  "inDefinedTermSet": "https://casrai.org/dictionary/domain/dictionary#set",
  "url": "https://casrai.org/dictionary/term/consort-flow-diagram",
  "sameAs": [],
  "license": "https://creativecommons.org/licenses/by/4.0/",
  "publisher": {
    "@id": "https://casrai.org/#organization"
  },
  "author": {
    "@id": "https://casrai.org/#editorial-team"
  },
  "datePublished": "2026-09-01T14:36:04",
  "dateModified": "2026-09-01T14:36:04",
  "inLanguage": "en-GB",
  "isAccessibleForFree": true
}