STROBE Checklist

What STROBE is

STROBE is the reporting guideline for observational studies in epidemiology. Its 22-item checklist covers the three main designs: cohort, case-control and cross-sectional studies. Most items are shared across all three, and a few have design-specific wording. Journals that publish clinical research generally expect STROBE for observational work, and many ask for the completed checklist at submission.

When it applies

Use STROBE when you observe exposures and outcomes without assigning an intervention. That includes retrospective chart reviews, registry analyses and database studies. Related extensions cover specific situations, notably RECORD for studies using routinely collected health data such as electronic records and claims.

What the checklist covers

• Title and abstract: the study design named with a commonly used term, and a balanced summary.
• Introduction: background, rationale and objectives, including any prespecified hypotheses.
• Methods: setting and dates, eligibility criteria and how participants were selected, clear definitions of exposures, outcomes, confounders and effect modifiers, data sources, efforts to address bias, how the study size was arrived at, and the statistical methods, including handling of missing data.
• Results: numbers at each stage of the study (a flow diagram is recommended), descriptive data, outcome data, and both unadjusted and adjusted estimates with their precision.
• Discussion: key results, limitations including the direction and size of likely bias, interpretation and generalisability.
• Other information: funding and the role of the funder.

What authors most often miss

• Confounding handled by naming covariates without explaining why they were chosen.
• Missing data not reported, or excluded silently from the analysis.
• Unadjusted estimates omitted, so readers cannot see how much adjustment changed the result.
• No account of how the study size was determined, even when it was simply all eligible records.
• Causal language in the conclusions that the design cannot support.

Ophthalmology: state your unit of analysis

Observational eye studies often include both eyes of some patients and one eye of others. Say whether the analysis is per eye or per patient, how the study eye was chosen when only one was used, and how you handled the correlation between fellow eyes. This is one of the most common methodological queries from ophthalmology reviewers.