Case-control studies for the evaluation of screening.

The case-control approach has been proposed for the evaluation of the efficacy of screening for cancer. It provides information at the level of the individual subject and allows direct estimation of the relative risk of not being screened. The criteria for the choice of the case and control series are examined, with reference to the type of cancer considered and the precise goal assigned to the screening programme. The definition of the exposure is discussed, as well as criteria for the avoidance of commonly occurring biases. Typical situations where case-control studies of screening efficacy appear to be the method of choice are presented.