9. Cohort Studies: Prospective, Retrospective, Open, and Closed
After this lesson you can
- Define a cohort study by its sampling and draw its schematic with time running forward
- Distinguish prospective from retrospective cohorts and tell a retrospective cohort from a case-control study
- Match closed cohorts to risks and open cohorts to rates
- Explain the healthy worker effect and why an internal comparison group is fairer
- Bound an estimate under loss to follow-up and plan follow-up longer than induction plus latency
- Compute and report the risk ratio, rate ratio and risk difference from a cohort table
Video coming soon
Scheduled for release on November 6, 2026.
This lesson teaches the cohort study, defined by sampling on exposure and following forward to outcome. It separates prospective from retrospective cohorts as a matter of timing, open from closed cohorts by the measure each yields, and walks through the four mechanics that threaten a cohort, using the invented Oakhaven factory cohort to compute and report the risk ratio, rate ratio and risk difference.
Questions and answers
Questions a learner would ask about this lesson. Think through your own answer before you open each one.
- What makes a study a cohort study?
Show answer
A cohort study samples people by exposure status, records whether each participant is exposed or not, and then tracks them to see who develops the outcome. Nobody is chosen because of their outcome, and everyone starts free of the outcome. The direction of inquiry runs from exposure forward to outcome.
- Is a retrospective cohort a case-control study, since both look at the past?
Show answer
No. The calendar does not decide the design; the sampling does. A retrospective cohort is sampled by exposure from existing records and its outcomes are traced forward in those records, while a case-control study is sampled by outcome and traces exposure back. Prospective and retrospective describe where the investigator stands in time, not two different designs.
- What is the difference between an open and a closed cohort?
Show answer
In a closed cohort membership is fixed at baseline and everyone is followed for one window, so it yields a risk: cases over people at the start. In an open, or dynamic, cohort members enter and leave during follow-up, so it yields a rate: cases over person-time. Computing a risk on an open cohort is an error.
- What is the healthy worker effect?
Show answer
Employed people are healthier on average than the general population, because the sick and disabled are less often hired or retained. So an occupational cohort compared with the general population can look protected even with a harmful exposure. In the lesson, the factory's death rate of 5 per 1,000 against the district's 8 per 1,000 gives a ratio of 0.63, which reflects selection, not protection; the fair comparison is internal, exposed against unexposed workers at the same factory.
- Why does the length of follow-up matter?
Show answer
A cohort must wait through the induction period, from exposure to the start of disease, and the latent period, from the start of disease until it can be detected. Follow-up shorter than induction plus latency finds nothing, and that null result is not reassurance. Plan follow-up from what is known about the condition and report how long the cohort watched.
Practice quiz
- Question 1What single rule separates a cohort study from a case-control study?
- Question 2Exposure and outcome have both already occurred, and an investigator reconstructs exposure groups from records and traces their outcomes. What is this?
- Question 3Membership is fixed at baseline and nobody joins after the start. Which frequency measure does this cohort yield?
- Question 4The factory's all-cause death rate is 5 per 1,000 and the district's is 8 per 1,000, a ratio of 0.63. What does that ratio mainly reflect?
- Question 5Of 400 exposed workers, 40 leave mid-year with unknown outcomes; 100 cases are seen. If 20 of the 40 lost workers had fallen ill, what would the exposed risk be?
- Question 6In the same cohort, what is the observed exposed risk among the 360 workers who were followed?
- Question 7What is the induction period?
- Question 8A cohort followed for less than induction plus latency finds no excess cases. How should that result be read?
- Question 9Which design is best for a rare exposure with several outcomes of interest?
- Question 10What does a nested case-control study do?
Saved in your browser only: no account, no server.
Principles of Epidemiology is education in epidemiologic reasoning, not clinical, screening, or policy guidance. Every example is set in the fictional Oakhaven Health District, and every number is invented for teaching.