Case control study
These compare the use of an intervention in groups with and without a particular condition. They can provide useful evidence on associations but need careful design to minimise selection bias and confounding. They are particularly useful in epidemiological studies of long term and low incidence effects of a therapy. They are often hypothesis generating, but may need confirmation by randomised controlled trial evidence before an effect of a therapy is confirmed to be significant. This happened with certain COX II inhibitors.
Appraising case-control studies
Case-Control studies are suitable for rare conditions, and can look at multiple exposures.
- Aims (and whether results appropriate to aims).
- Case definition
- Case ascertainment
- Control selection (?bias) - patients within the same medical facility, excluding those with certain diseases, including only those with certain diseased believed to be unrelated to the causal factors to be investigated…, sample of general community…)
- Exposure definition
- Exposure ascertainment
- Non-response/exclusions - are those analysed representative?
- Adjustments for problems - are they adequate?
- Outcome measurement.
- Bias - recall, recording, etc.
- Confounding (age, sex, socio-economic group, smoking, alchohol, vitamins, nutrition…) adequately considered?
This article is a work in progress. Please feel free to contribute to it.