Cohort Study: Definition and Example

A cohort study follows a group of people over time, classifying them by exposure at the start and comparing how often an outcome occurs in each exposure group. Because exposure is recorded before the outcome, cohort studies can estimate incidence and risk and show temporal order.

Types

• Prospective cohort: the investigators enrol participants and follow them forward.
• Retrospective (historical) cohort: the investigators use existing records to reconstruct exposure and outcome that have already occurred.
• Single or multiple exposures: a cohort can examine one exposure or many.

Key steps

1. Define the population, exposure and outcome precisely.
2. Recruit or identify exposed and unexposed groups from comparable sources.
3. Measure confounders at baseline.
4. Follow up for a defined time with outcome ascertainment that does not depend on exposure status.
5. Analyse risk ratios, rate ratios or hazard ratios with confidence intervals, adjusting for confounders.
6. Report losses to follow-up and following STROBE.

Illustrative example

Invented for teaching. In a retrospective cohort, 1,000 patients with a hip fracture are classified by whether they had surgery within 48 hours. Delirium occurs in 90 of 400 patients operated within 48 hours (22.5%) and 210 of 600 operated later (35%). The risk ratio for early surgery versus later is 0.64. Because patients operated later may have been sicker, the result needs adjustment for confounding.

Strengths and limitations

Strengths: temporal sequence, multiple outcomes, incidence estimates. Limitations: time and cost (prospective), loss to follow-up, confounding, and misclassification in routine records.

Common pitfalls

• Immortal time bias (follow-up time during which the outcome cannot occur is counted in the exposed group).
• Differential follow-up between groups.

Key takeaways

• Cohorts follow exposure to outcome and estimate risk or hazard.
• Careful control of confounding and attrition is essential.

This page is general educational information about research methods. It is not medical, legal or regulatory advice; follow your institution's ethics board and your target journal's instructions.