CONSORT, STROBE, PRISMA or CARE? How to choose the right reporting guideline

Journals increasingly ask authors to submit a completed reporting checklist with their manuscript, and editors may return a paper without review if it is missing. The checklist you need depends almost entirely on one thing: your study design. Get the design right and the guideline follows.

This guide explains what reporting guidelines are, gives a decision guide for the common designs in clinical research, and lists the extensions that catch many authors out.

What a reporting guideline does

A reporting guideline is a checklist, usually with a flow diagram, that specifies the minimum information readers need to understand how a study was done and to judge whether its results are trustworthy. Guidelines are developed through consensus among methodologists, journal editors and researchers. The EQUATOR Network (Enhancing the QUAlity and Transparency Of health Research) hosts a searchable library of them.

Guidelines improve reporting, not research. They cannot rescue a poorly designed study, but they ensure that a well-designed one is not undersold, and they make peer review faster because reviewers can find what they are looking for.

Step 1: name your study design precisely

Before opening any checklist, write one sentence describing the design, such as "a prospective cohort study of 240 patients undergoing elective surgery" or "a randomised, double-blind, placebo-controlled trial". Ask yourself:

• Did the investigators assign the intervention? If yes and allocation was random, it is a randomised trial.
• If the investigators only observed, were participants followed over time (cohort), selected by outcome (case-control) or assessed at one point (cross-sectional)?
• Are you evaluating how well a test identifies a condition? That is a diagnostic accuracy study.
• Are you summarising other studies systematically? That is a systematic review, with or without meta-analysis.
• Are you describing one patient or a few? That is a case report or case series.

Step 2: match the design to the guideline

Step 3: check for an extension

Many core guidelines have extensions for particular situations, and journals expect the extension when it applies. Examples include CONSORT extensions for cluster randomised and non-inferiority trials, STROBE extensions for molecular epidemiology (STROBE-ME) and for routinely collected health data (RECORD), PRISMA extensions for scoping reviews (PRISMA-ScR) and network meta-analyses, and reporting guidance for abstracts (CONSORT for Abstracts, PRISMA for Abstracts). Search the EQUATOR library for your design plus a keyword before you start.

Step 4: use the checklist while writing, not after

The checklist is most useful as a writing outline. Draft the methods section item by item, then record the page or line number for each item as you go. Items that you cannot complete reveal real gaps: perhaps the sample-size calculation was never documented, or the handling of missing data was never decided. It is far better to discover that before submission than in a reviewer's report.

A useful test: give your methods section and the checklist to a colleague who was not involved in the study. Every item they cannot find is an item a reviewer will query.

Common mistakes

• Using CONSORT for a non-randomised comparison. Non-randomised intervention studies have their own guidance (for example TREND); observational comparisons use STROBE.
• Claiming to follow a guideline without reporting its key items, such as the flow diagram numbers.
• Completing the checklist with "yes" instead of page numbers, which tells the editor nothing.
• Ignoring the abstract extension, even though the abstract is what most readers see.

Sources and further reading

• EQUATOR Network library of reporting guidelines
• ICMJE Recommendations, which encourage the use of reporting guidelines

Guidelines are periodically updated. Always download the current checklist from the guideline's official website or the EQUATOR library.

If you would like a second pair of eyes, our reporting guideline review checks the manuscript item by item and returns an annotated checklist in about three working days.