Case-control study
Why this level?
Specify the research question: intervention, diagnostic accuracy, prognosis, aetiology, or screening intervention. Other question types are outside this hierarchy. NHMRC 2009 sourceIntervention case-control studies with concurrent controls and aetiology case-control studies occupy different categories. For aetiology the same sampling family has a different level. Selection of controls and the timing of exposure ascertainment still need careful appraisal.
Where the NHMRC 2009 hierarchy names this design, it places it here:
| Research question | Level | When |
|---|---|---|
| Intervention | III-2 | An observational comparison uses concurrent controls. Statistical adjustment does not change this design level. |
| Aetiology | III-3 | A case-control study examines the exposure in people selected by outcome status. |
What question does this help answer?
“Were people with this outcome more likely than appropriate controls to have had a particular prior exposure?”
When to use it
Case-control designs are especially useful for rare outcomes, long-latency diseases, or questions where assembling a full cohort would be inefficient.
When not to use it
It is usually a poor choice when the exposure itself is rare, when valid controls cannot be drawn from the same source population as the cases, or when exposure information is highly vulnerable to differential recall or recording.
How it works in plain language
Cases are selected because they have the outcome. Controls should represent the exposure distribution that would have been observed in the population that produced the cases. Researchers then compare previous exposure between cases and controls, usually estimating an odds ratio.
What the sampling looks like
What a Table 1 looks like, by case status
| Characteristic | Group 1 (n = 250) | Group 2 (n = 250) | P value |
|---|---|---|---|
| Age, mean (SD), y | 64.2 (12.1) | 63.0 (13.1) | .07 |
| Female, No. (%) | 142 (56.8) | 138 (55.2) | .74 |
| Comorbidity index, mean (SD) | 2.6 (1.7) | 2.5 (1.9) | .54 |
| Insurance type, No. (%) | 62 (27.2) | 56 (25.2) | .64 |
| Baseline severity score, mean (SD) | 5.4 (2.2) | 5.2 (2.4) | .75 |
Why Mallard may recommend it
Mallard may recommend a case-control design when the outcome is uncommon and the user has a defensible source population and reliable historical exposure data.
Key assumptions and design requirements
- Cases and controls come from the same underlying source population.
- Control selection is independent of exposure except through appropriate sampling.
- Exposure is measured comparably in cases and controls.
- Confounders are identified using design and subject-matter knowledge.
- Matching, if used, is reflected correctly in analysis.
What to watch out for
Poor control selection can invalidate the study. Recall bias, selection bias, overmatching, and interpreting an odds ratio as if it were always a risk ratio are common problems.
Clinical example
Identify patients with a rare postoperative complication and controls from patients who underwent the same procedure without the complication, then compare prior exposure to a medication.
- STROBE statement for observational studies (opens in a new tab) · EQUATOR Network
- Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: guidelines for reporting observational studies (opens in a new tab) · BMJ · doi 10.1136/bmj.39335.541782.AD
- Case-control studies: Increasing scientific rigor in control selection (opens in a new tab) · Res Pract Thromb Haemost · doi 10.1016/j.rpth.2023.100090
- Observational studies: cohort and case-control studies (opens in a new tab) · Plast Reconstr Surg · doi 10.1097/PRS.0b013e3181f44abc