Which reporting guideline does my study need?

What is the EQUATOR Network?

The EQUATOR Network (Enhancing the QUAlity and Transparency Of health Research) is an international initiative that collects and promotes reporting guidelines for health research. Its online library is the standard place to find the right checklist. As of October 2026 it lists more than 700 reporting guidelines, from the major ones such as CONSORT and PRISMA to narrow extensions for single methods or specialties.

A reporting guideline is a minimum list of what a paper must tell readers so they can understand, appraise and repeat a study. It usually comes as a checklist, often with a flow diagram and an explanation-and-elaboration paper. Reporting guidelines describe what to report, not how to run the study. They are not tools for scoring study quality.

Which reporting guideline does my study need?

Start from what you actually did, not from what the paper is called. This table matches common clinical study designs to the main guideline listed by EQUATOR, with its current version as of October 2026.

If your design is not in the table, search the EQUATOR library by design and by specialty. Many guidelines exist for narrow methods, such as Mendelian randomization (STROBE-MR) or nutritional epidemiology (STROBE-nut).

How do I choose when my study fits more than one guideline?

1. Name the design in one sentence. Did you assign the intervention (a trial), observe exposures you did not control (observational), pool published studies (a systematic review), or describe one or a few patients (a case report or series)?
2. Pick the main guideline for that design from the table above. You only ever have one main guideline.
3. Add extensions for your methods or data. Extensions are used together with the main checklist. Examples: a retrospective EHR cohort uses STROBE plus RECORD; a cluster-randomised trial uses CONSORT plus its cluster extension; a meta-analysis of prediction models uses PRISMA plus the relevant TRIPOD extension.
4. Use the protocol version for protocols. SPIRIT and PRISMA-P are for protocols. CONSORT and PRISMA are for completed studies.
5. Check for an abstract checklist. CONSORT, PRISMA and STROBE all have versions for abstracts. Some journals and meetings ask you to follow them.
6. Confirm with the journal. If the instructions for authors name a different checklist for your article type, follow the journal.

EQUATOR also publishes a simple flow chart that helps match designs to guidelines, on its page on selecting the appropriate reporting guideline.

What about AI and machine-learning studies?

AI studies now have their own guidelines, and reviewers expect them. If you develop or validate a model that predicts a diagnosis or outcome, use TRIPOD+AI. It replaces TRIPOD 2015 for both regression and machine-learning models. If you evaluate a large language model, TRIPOD-LLM extends TRIPOD+AI. Diagnostic accuracy studies of an AI test use STARD-AI. An early live clinical evaluation uses DECIDE-AI. A randomised trial of an AI intervention uses CONSORT with its AI extension.

This is separate from disclosing your own use of AI to write the paper. Journals have their own policies on that, and authors remain responsible for every statement. See our guide to the AI disclosure statement for medical journals.

Which journals require a reporting checklist?

Requirements vary by journal and by article type, so read the instructions for authors for your target journal. Here is what some key policy sources say as of October 2026:

Many journals ask for the completed checklist at submission, and some return a submission without it before review. Some also ask you to state in the cover letter which guideline you followed. If your journal asks for this, add one sentence to your cover letter for journal submission.

How do I submit the checklist with page or line numbers?

1. Download the official checklist from the guideline's website or the EQUATOR library entry. Use the version for your design (for example, the cohort version of STROBE).
2. Turn on continuous line numbers in the final manuscript, so references stay stable.
3. For every item and sub-item, write where it is reported: page and line numbers, or a table, figure or supplement reference. Some journals also ask you to paste the sentence that addresses the item.
4. Write "not applicable" with a short reason for items that do not apply. Never leave a row blank.
5. Fix the manuscript, not the checklist. If an item is missing, add the information to the paper.
6. State the guideline in the Methods (for example, "reported following the PRISMA 2020 statement") and cite its publication.
7. Upload the checklist as the journal's specified file type, usually a supplementary file or "reporting checklist".
8. Update the line numbers after each revision, and resubmit the updated checklist with your response to reviewer comments.

Worked example: choosing guidelines for two studies (synthetic)

Both studies below are made up for illustration. They do not describe real research.

Study A. A team reviews hospital records from 2019 to 2024 and compares postpartum haemorrhage rates between women induced with two different protocols. Nobody assigned the protocol for research purposes, so this is a retrospective cohort study. The guideline is the STROBE checklist for observational studies, plus RECORD because the data are routinely collected records. A checklist row might read: "STROBE 6a / RECORD 6.1: Methods, p. 5, lines 88–97 (eligibility, data source and codes used to identify the cohort)."

Study B. The same team then randomises 240 women to the two protocols. This is a randomised trial. The protocol follows SPIRIT 2025 and the results paper follows the CONSORT 2025 checklist and flow diagram. If they later pool their trial with others in a meta-analysis, that paper follows PRISMA 2020 and its PRISMA flow diagram.

Common mistakes reviewers flag

• The wrong guideline for the design, for example CONSORT for a non-randomised comparison or STROBE for a case series of three patients. A single patient needs CARE: see how to write a case report.
• An outdated version, such as CONSORT 2010 or TRIPOD 2015, when CONSORT 2025 and TRIPOD+AI are current.
• Missing extensions, such as RECORD for a registry study or STROBE-MR for a Mendelian randomization paper.
• A checklist with "Methods" in every row, or page numbers from an earlier draft.
• Claiming compliance in the Methods while key items are clearly missing, such as the participant flow or how missing data were handled.
• Treating the checklist as an appraisal score in a systematic review, instead of using a risk-of-bias tool.

CliniPaper identifies the design of your manuscript, suggests the matching guideline and extensions, and checks the text item by item. You still need to confirm the final choice against your target journal's instructions.