5.2 CONSORT 2025 & Flow Diagrams for Randomized Trials

CONSORT 2025: Randomized Trial Reporting

Purpose and Scope

CONSORT helps authors report randomized trials with enough clarity for readers to understand what was planned, what was done, what was found, and how the findings should be interpreted. It does not certify that a trial was well designed, and checking every box cannot repair poor methods. It also does not replace the protocol, statistical analysis plan (SAP), trial registration, ethics documentation, or journal instructions.

The central publication task is traceability:

registered question → protocol → SAP → conduct → analyzed data → reported results

Differences can occur, but important changes must be identified and explained rather than hidden.

Title, Abstract, and Introduction

Identify the report as a randomized trial in the title so readers and indexers can find it. The abstract should use the appropriate structured reporting guidance and accurately summarize design, participants, interventions, prespecified outcomes, major results, harms, registration, and funding within the journal's limits.

The introduction explains the scientific background and rationale and states specific objectives or hypotheses. Avoid retrospective claims that make an exploratory analysis appear prespecified.

Trial Design and Protocol Changes

Report the design precisely: parallel, crossover, cluster, factorial, noninferiority, superiority, adaptive, or another form. State the allocation ratio and relevant framework. If methods changed after the trial began, identify the change, when it occurred, and why.

A publication professional should reconcile at least four sources before drafting:

1. registry record;
2. approved protocol and amendments;
3. final SAP;
4. clinical database and analysis outputs.

A mismatch is not automatically misconduct. The ethical problem is presenting the final report as if a material change never happened.

Participants, Settings, and Interventions

Describe eligibility criteria and the settings and locations where data were collected. Intervention descriptions should be detailed enough to understand what each group received and, where feasible, to support replication. This includes timing, dose or intensity, delivery, permitted co-interventions, and relevant adherence information.

For behavioral or complex interventions, use the applicable extension or intervention-description guidance rather than forcing every detail into an oversimplified drug-trial template.

Outcomes and Harms

Define prespecified primary and secondary outcomes clearly, including the measurement variable, analysis metric, aggregation method, and time point where relevant. Report changes to outcomes after trial commencement with reasons. Do not silently replace an unfavorable prespecified outcome with a favorable exploratory one.

Harms deserve planned, balanced reporting. Describe how adverse events were defined, collected, analyzed, and presented. A statement that an intervention was “well tolerated” is not an adequate substitute for data.

Sample Size and Early Stopping

Explain how sample size was determined, including assumptions needed to understand the calculation. If interim analyses or stopping guidelines were used, describe them. If enrollment ended early or the trial stopped, report who made the decision and why. Readers need to distinguish a planned information-based stopping rule from an operational or commercial decision.

Randomization: Three Separate Questions

A computer-generated sequence can still be undermined if recruiters can predict the next assignment. Conversely, sealed envelopes do not rescue a patterned or nonrandom sequence. Report both sequence generation and the concealment mechanism.

Masking

State who was masked after assignment—participants, care providers, outcome assessors, data analysts, or others—and how masking was maintained, if applicable. Avoid “single-blind” and “double-blind” without naming the groups.

If masking was impossible, say so and describe safeguards against bias. Do not claim interventions were identical unless the relevant similarities are explained.

Participant Flow and Numbers Analyzed

The flow diagram should reconcile the path from assessment and randomization through allocation, follow-up, and analysis. For each group, report numbers receiving the assigned intervention, losses and exclusions, and reasons.

The denominator must be clear for every analysis. The phrase “intention-to-treat” should not conceal exclusions, missing data, or treatment switching. State which participants were included and how missing outcome data were handled. If additional per-protocol or as-treated analyses are shown, label and justify them.

Results and Estimation

Report recruitment and follow-up dates, baseline information, numbers analyzed, and results for each prespecified outcome. For each group, give the result and an effect estimate with precision, ordinarily a confidence interval. For binary outcomes, absolute and relative effects may both help interpretation.

Separate prespecified subgroup or adjusted analyses from exploratory work. Report all important harms and unintended effects. Statistical significance alone does not establish clinical importance; interpret magnitude, precision, multiplicity, consistency, and limitations.

Example

If 240 participants were randomized 1:1, the report should not simply state that 220 completed the study. It should show:

• 120 allocated to each group;
• how many received the assigned intervention;
• losses and reasons by group;
• analysis denominators by outcome;
• handling of missing data;
• effect estimates and confidence intervals.

This allows readers to judge attrition and analysis choices.

Discussion and Other Information

Discuss limitations, potential bias, imprecision, multiplicity, and generalizability. Interpret benefits and harms together and in the context of other evidence. Avoid causal claims beyond the randomized contrast studied.

Provide registration information and access information for the protocol, SAP, data, code, or other materials as required by CONSORT 2025, the journal, funder, and applicable policy. State funding and the funder's role. Open-science statements must be truthful: a data-sharing statement may explain that data are unavailable, but it must not promise access the sponsor cannot provide.

Publication Workflow

Before submission:

1. download the current CONSORT 2025 checklist from the official CONSORT-SPIRIT site;
2. identify any applicable extension;
3. map each item to a manuscript page or explain non-applicability;
4. reconcile outcomes and analyses to the registry, protocol, and SAP;
5. verify flow counts against tables and outputs;
6. confirm author approval and disclosure statements;
7. recheck target-journal instructions.