Mortality in Relation to Smoking: Ten Years' Observations of British Doctors

Abstract

In previous papers (Doll and Hill, 1954, 1956) we have described how at the end of October 1951 we sent a short and simple questionary to the 59,600 men and women whose names were on the current British Medical Register and who were then resident in the United Kingdom.In addition to giving name, address, and age, they were asked to say whether (a) they were, at that time, smokers of tobacco, (b) they had previously smoked but had given up, or (c) they had never smoked regularly (which we defined as having never smoked as much as one cigarette a day, or its equivalent in pipe tobacco or cigars, for as long as one year).The smokers and ex-smokers were asked the age at which they had started smoking, the amount that they smoked, and the method by which they smoked either at the time of reply or when they last gave up, and, when appropriate, the age at which they had stopped.We deliberately limited our inquiries to these very few ques- tions, partly to encourage a large number of answers and partly because we believed that these were questions that could be answered with reasonable accuracy.For such reasons we did not ask for a life-history of smoking habits nor did we, at that time, inquire into the habit of inhaling. DataTo this request we had 40,637 replies sufficiently complete to be used-34,445 from men and 6,192 from women.From a 1 in 10 random sample of the register that we subsequently drew and analysed we estimate that these figures represent answers from 69% of the men and 60% of the women alive at the time of the inquiry.' (These numbers differ slightly from those we published in 1956; from answers to our second questionary (see below) we learned that we had allotted the wrong sex to a few subjects and had included a few forms that had come from relatives who were not doctors.) Selective SampleWe may feel sure that the doctors who chose to answer were not representative of the total.The seriously ill would have been unable to respond, and thus, as we showed in our previous paper, the mortality of the group who replied would be, at least for a time, abnormally low.In fact, using the 1 in 10 sample as a basis, we calculated that the standardized

Keywords

Related Publications