[Durgical](https://durgical.com/en/) › [DurgiPedia](https://durgical.com/en/durgipedia/) › [Epidemiology](https://durgical.com/en/epidemiology/) › [Study designs](https://durgical.com/en/epidemiology/areas/designs/) › Case-cohort study

# Case-cohort study

- Study design

Inside a large cohort, all cases plus a random subcohort chosen at the start are analysed; the subcohort serves as comparison for several diseases.

Updated Nov 13, 2026

## Key points

- Random subcohort at baseline + all cases from follow-up.
- The subcohort is chosen without knowing who will fall ill; it can include future cases.
- One subcohort can be used to study several diseases.
- In a nested case-control study, controls are drawn from those still at risk at each case's time and are usually matched.
- Analysed with weighted Cox models that correct for the over-representation of cases.

On this page

- The basics Key points
- What it is
- How it is built
- Worked example
- Case-cohort versus nested case-control and full cohort
- Markers
- When to use it
- Common pitfalls
- Memory aid

- Study Test yourself

- Connections Connections

## What it is

A case-cohort study is built inside a large cohort with stored samples or data. At baseline a random subcohort is drawn that represents the whole cohort; during follow-up every case that arises is added, whether or not it is in the subcohort. Expensive exposures are measured only in this sample, and the same subcohort serves as the comparison group for several different outcomes.

## How it is built

- Start from a defined cohort with biological samples or data stored at baseline.
- Draw a random (simple or stratified) fraction of the cohort: the subcohort.
- Follow the whole cohort and identify every case of each outcome.
- Measure the expensive exposures only in the subcohort and in the cases outside it.
- Estimate hazard ratios with weighted methods.

## Worked example

- Cohort of 12 000 workers with blood frozen at baseline. A random 5 % subcohort is drawn: 600 people.
- Over 8 years, 180 cases of diabetes arise; 9 of them were in the subcohort (5 % of 180 = 9, as expected).
- Samples to analyse for diabetes: 600 + (180 − 9) = 771, instead of 12 000.
- Stroke is then studied with the same subcohort: 90 cases, 5 inside it; 90 − 5 = 85 samples are added.
- Total for two diseases: 600 + 171 + 85 = 856 samples, about 7 % of the cohort (856 / 12 000 = 7.1 %).
- Reading: two outcomes are studied by measuring the biomarker in fewer than 1 in 14 participants.

## Case-cohort versus nested case-control and full cohort

| Aspect | Case-cohort | Nested case-control | Full cohort |
| --- | --- | --- | --- |
| Comparison group | Random subcohort chosen at baseline | Controls chosen at each case's time | All non-cases |
| Several outcomes | Yes, with the same subcohort | New controls needed for each | Yes |
| Matching | No (adjusted in the analysis) | Usual (age, time in cohort) | Not needed |
| Samples to analyse | Subcohort + cases | Cases + controls | All |
| Analysis | Weighted Cox | Conditional logistic | Usual Cox or rates |

## Markers

- Subcohort
- HR
- Weighted Cox

## When to use it

- Biobanks and large cohorts in which measuring a biomarker or gene in everyone would be very costly.
- When several outcomes are to be studied with the same comparison group.
- It also allows absolute risks to be estimated, because the subcohort represents the full cohort.

## Common pitfalls

- Analysing it as an ordinary case-control study: ignores that the subcohort contains future cases and follow-up time.
- Choosing the subcohort after knowing who fell ill: it is no longer a random sample of the cohort.
- Using it for a very common outcome: cases would be nearly the whole cohort and little is saved.

## Memory aid

Case-cohort: one random ‘snapshot’ of the cohort at baseline, reused with every new disease.

## Test yourself

Think of the answer, then tap the question to see it.

## What is the key difference between a case-cohort and a nested case-control study?

In a case-cohort the comparison is a random subcohort drawn at baseline; in a nested study controls are picked at each case's time from those still at risk.

## Can a subcohort member later become a case?

Yes. The subcohort is chosen without regard to outcome, so some members will fall ill and count as cases.

## What advantage does a case-cohort design have when several diseases are studied?

The same subcohort serves as comparison for all of them; only the new cases of each outcome are added.

[Add to my cards (3)](https://durgical.app/estudiar/?add=/en/epidemiology/case-cohort-study/)

## Connections

## Related 5

- [Nested case-control study](https://durgical.com/en/epidemiology/nested-case-control-study/)
- [Cohort study](https://durgical.com/en/epidemiology/cohort-study/)
- [Case-control study](https://durgical.com/en/epidemiology/case-control-study/)
- [Hazard ratio and the Cox model](https://durgical.com/en/epidemiology/hazard-ratio-and-the-cox-model/)
- [Survival analysis and censoring](https://durgical.com/en/epidemiology/survival-analysis-and-censoring/)

[← All of epidemiology](https://durgical.com/en/epidemiology/)

Educational content on epidemiological and statistical methods; the examples use made-up numbers to teach the calculation. Not a clinical or legal guide: current Colombian rules are in the official sources.

Sources: [MrOS case-cohort data](https://agingresearchbiobank.nia.nih.gov/studies/mros/documents/download/Case_Cohort_Overview.pdf/) (National Institute on Aging, Aging Research Biobank, 2012); [Principles of Epidemiology | Lesson 1 - Section 7](https://archive.cdc.gov/www_cdc_gov/csels/dsepd/ss1978/lesson1/section7.html) (Centers for Disease Control and Prevention, 2012); [Chapter 25: Assessing risk of bias in a non-randomized study](https://www.cochrane.org/authors/handbooks-and-manuals/handbook/current/chapter-25) (Cochrane, 2024).

## **Spotted a mistake?** We are always improving for the community. Tell us what is wrong and our team will review it. Report a mistake

Representative model, for educational purposes only.

---
Source: https://durgical.com/en/epidemiology/case-cohort-study/
License: CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). Cite as: DurgiPedia (durgical.com), https://durgical.com/en/epidemiology/case-cohort-study/