Case-control studies of screening procedures

Case-control studies of different screening procedures have been reported. Cases of the detectable disease, and controls without the disease, were compared with respect to their previous screening histories. Measures of the protective value of the test were calculated from the difference. The inductive basis of these studies is re-examined and it is concluded that all claims to have measured the protective value of a screening test through such means are potentially or actually flawed. The reasons for the inductive failures are: (a) The ratio between the retrospective screening frequencies depends critically upon unmeasured parameters other than procedural efficacy; the ratio is therefore not capable on its own of providing an estimate of the protection afforded: (b) Almost all screens among reported case and control groups have been negative, and it is a matter of prior fact that a negative test does not protect anyone. Only the positive tests protect. Direct analogies with drug and vaccine trials, from which the concept of a 'protective value' originates, are therefore false.