From Exposure to Outcome: Building a Credible Cohort Study

Core logic of a cohort study

A cohort study defines a population, measures exposures and evaluates the occurrence of outcomes over time. Participants should be selected by eligibility and exposure information rather than by future outcome status. Temporality is stronger than in a cross-sectional design, but observational data remain vulnerable to confounding.

Prospective versus retrospective

Prospective cohorts establish the protocol before follow-up, allowing greater control over measurement at the cost of time and resources. Retrospective cohorts reconstruct exposure and follow-up from existing records. They can be efficient but are constrained by data quality, missing variables and historical changes in practice.

Align time zero

Eligibility, exposure assignment and start of follow-up should be aligned. Misalignment can generate immortal-time bias. Define the source population, entry date, baseline outcome status and clinically meaningful eligibility criteria explicitly.

Measurement and follow-up

Use reproducible exposure definitions, account for time-varying exposures when relevant and ascertain outcomes with standard criteria. Report attrition by group and investigate whether loss to follow-up is related to exposure and outcome risk.

Confounding

Do not select confounders solely by univariable p values. Use clinical knowledge, prior evidence and causal diagrams. Avoid adjusting for mediators or colliders without a clear estimand.

Common biases

• Selection bias and healthy-user effects

• Exposure or outcome misclassification

• Immortal-time bias

• Time-varying confounding

• Informative loss to follow-up

• Missing data and changes in record systems

Reporting

The STROBE checklist supports transparent reporting of participant flow, variables, bias, study size, missing data, follow-up and sensitivity analyses. It is a reporting guideline, not a quality-scoring instrument.

References

• EQUATOR Network: STROBE Statement.

• von Elm E et al. STROBE Statement. BMJ 2007.

Accessed 20 June 2026.