STROBE (Strengthening the Reporting of Observational Studies in Epidemiology) is the reporting guideline for the three main observational designs: cohort, case-control and cross-sectional studies. Most medical journals ask authors to submit the completed checklist, giving the page on which each item appears. It does not tell you how to do the study; it lists what the paper must say so that readers can judge it.
The structure
The 22 items follow the paper's sections. Eighteen apply to every design; four (items 6, 12, 14 and 15) have versions for cohort, case-control and cross-sectional studies.
| Section | Items | What they ask for |
|---|---|---|
| Title and abstract | 1 | The design in the title or abstract, and an informative, balanced abstract |
| Introduction | 2-3 | Background and rationale; specific objectives and hypotheses |
| Methods | 4-12 | Design, setting and dates, participants and how they were selected, variables (outcomes, exposures, confounders), data sources and measurement, bias, study size, handling of quantitative variables, statistical methods |
| Results | 13-17 | Participants at each stage (ideally a flow diagram), descriptive data, outcome data, main results with confounder-adjusted estimates and CIs, other analyses |
| Discussion | 18-21 | Key results, limitations, interpretation, generalisability |
| Other | 22 | Funding |
The items reviewers most often find missing
- Item 9, bias: what you did to address possible sources of bias, not only a list of them in the limitations.
- Item 10, study size: how the sample size was arrived at. See how to calculate sample size for a survey.
- Item 12, statistical methods: how confounders were controlled, how missing data were handled, and any sensitivity analyses.
- Item 13, participants: numbers at each stage (eligible, examined, included, analysed) and reasons for non-participation.
- Item 14, descriptive data: participant characteristics and missing data per variable, usually as Table 1.
- Item 16, main results: unadjusted and adjusted estimates with 95% CIs, and which confounders were adjusted for and why.
How to fill it in
- Download the checklist for your design from the STROBE website.
- For each item, write the page (and paragraph) where it is reported, not "yes".
- If an item does not apply, say why ("not applicable: no follow-up").
- Use the gaps as a to-do list before submission: an empty item is a likely reviewer comment.
Mistakes reviewers catch
- A checklist that says "yes" everywhere with no page numbers.
- Unadjusted results only, or adjusted results without saying for what.
- Missing data not reported for each variable.
- Causal language ("X causes Y") in a cross-sectional study.
Let KyroStat do the write-up for you
KyroStat produces the parts of a STROBE-compliant Results section that come from the data: participant flow numbers, Table 1 with missing data, and unadjusted and adjusted estimates with confidence intervals.
Frequently asked questions
Which guideline applies to randomized trials? CONSORT. See the CONSORT 2025 guide.
Is there a STROBE version for my field? Yes, several extensions exist (for example STROBE-MR for Mendelian randomisation and RECORD for routinely collected data). The EQUATOR Network lists them.
How should the statistics be formatted? By your journal's style; most medical journals use AMA style. See how to report statistics in AMA style.