Strawberry pickers' foot drop.

the examination that detected the endometrial cancer. Unless the distortions of "prescriber bias" and "detection bias" are eliminated casecontrol studies become an act of epidemiologic faith, not an exercise in aetiological science. While claiming that "each and every" infidel offers "invalid" criticism, emanating from "misunderstanding," you display the same intellectual infirmities ascribed to the critics. Your scholarship is inxalid because it is incomplete and misleading; no citation or explanation is provided for the contradictory results of two previous case-control studies1 2 in which no relationship was found between oestrogens and endometrial cancer. You also suffer from a misunderstanding of the way that common clinical events can create prescriber bias and detection bias to produce a fallacious relationship between oestrogens and endometrial cancer. In one group of women, bleeding with undetectable or undetected early endometrial cancer, the oestrogen prescribed as treatment for the bleeding will be associated with the cancer when it later reaches a detectable state. In a second group of women, who have asymptomatic endometrial cancer, the oestrogens given for postmenopausal symptoms may provoke bleeding that leads to the diagnostic curettage. Both of these important confounding phenomena have received no attention in the case-control studies cited by