Case-Cohort Design
Case-cohort design is an epidemiological study design developed by Prentice (1986) that efficiently combines features of case-control and cohort studies. Researchers enroll an entire cohort, follow it for outcomes, then measure exposures only on cases and a random subcohort, reducing measurement costs while maintaining valid causal inference.
Read the full method
Sign in with a free account to read this section.
Method map
The neighbourhood of related methods — select a node to explore.
When to use it
Apply case-cohort design when exposure measurement is expensive or requires biological samples, when diseases are rare, or when long follow-up creates costs. Ideal for cohort studies of biomarkers, genes, or rare diseases. Provides valid causal estimates with lower cost than full cohort studies.
Strengths & limitations
- Cost-efficient: measures expensive exposures on only a fraction of cohort
- Maintains cohort structure: preserves longitudinal design and prospective exposure ascertainment
- Valid inference: specialized analysis methods account for sampling to provide unbiased estimates
- Handles rare diseases: can be more efficient than traditional case-control studies
- Flexible subcohort: can stratify subcohort by demographic or risk factors
- Statistical power: smaller exposure sample than full cohort reduces power for rare exposures
- Complex analysis: requires specialized software and understanding of weighted regression
- Measurement assumptions: must measure exposures in subcohort at same time as cases (not retrospectively)
- Potential bias: if subcohort lost to follow-up or exposures measured differently
Frequently asked
What is a subcohort and how do I select it?
The subcohort is a random sample of the cohort selected at baseline, before knowing outcomes. Use stratified random sampling: randomly select fixed number or proportion from each stratum (age, sex, baseline risk). Stratification can improve efficiency.
How do I analyze case-cohort data?
Use specialized methods: Prentice estimator for risk differences, Cox regression with weighted likelihood, or conditional logistic regression. Standard cohort analysis biases estimates. Software (R packages, SAS PHREG) handles case-cohort designs.
Can I use case-cohort if exposures weren't measured in subcohort at baseline?
This creates problems. Exposures should be measured prospectively in subcohort members at the same time as cases. Retrospective or inconsistent measurement biases results.
What is the trade-off between subcohort size and power?
Larger subcohorts have more power and allow studying more exposures, but increase costs. Typical subcohorts include 5-20% of cohort. Optimization depends on disease and exposure rarity.
How does case-cohort relate to case-control studies?
Similar efficiency but different design: case-control selects controls after outcome is known; case-cohort pre-specifies subcohort. Case-cohort is preferred for prospective exposure measurement.
Sources
- Prentice, R. L. (1986). A case-cohort design for epidemiologic cohort studies and disease prevention trials. Biometrika, 73(1), 1-11. DOI: 10.1093/biomet/73.1.1 ↗
- Barlow, W. E., Ichikawa, L., Rosner, D., & Izumi, S. (1999). Analysis of case-cohort designs. Journal of Clinical Epidemiology, 52(12), 1165-1172. DOI: 10.1016/S0895-4356(99)00102-X ↗
- Kang, S., & Cai, J. (2009). Spatial matched-pair cohort studies. Biometrics, 65(2), 526-534. link ↗
How to cite this page
ScholarGate. (2026, June 3). Case-Cohort Design. ScholarGate. https://scholargate.app/en/psychometrics/case-cohort-design
Which method?
Set this method beside its closest kin and read them side by side — the library lays the books on the table; the choice is yours.
- Latent Transition AnalysisPsychometrics↔ compare
- Partial Least Squares Structural Equation ModelingPsychometrics↔ compare
- Rule Space MethodologyPsychometrics↔ compare
- Value-Added ModelingPsychometrics↔ compare