University of Minnesota

British Doctors Study

Study Category: The Cohort Studies (1947-1972)
Years: 1951 - 2001
Location: United Kingdom
Principal Investigator: Doll, Richard


In the 1950s, retrospective cohort studies of smoking status and lung cancer outcomes showed “more heavy smokers and fewer non-smokers among patients with lung cancer than among patients with other diseases.” [1] While the retrospective studies found a strong association between heavy smoking and lung cancer prevalence etc., the investigators sought to conduct a prospective study to “determine the frequency with which the disease appeared, in the future, among groups of persons whose smoking habits were already known.” [1] In addition to cancers, among other outcomes of interest were coronary thrombosis and other cardiovascular diseases.


The investigators chose UK practicing panel doctors as their population of interest and in October, 1951 mailed them questionnaires on demographics, smoking status, and medical history. Of the 59,600 questionnaires mailed, 41,024 were returned and 40,564 were properly completed. Because 10,017 questionnaires were returned by men aged less than 35 years and women of all ages, they were excluded due to the rarity of lung cancer in these cohorts. [1]


The association between smoking and incidence of lung cancer and coronary disease was examined in the men of the cohort after 29 months follow-up and for both diseases was highly significant and strongly related to the amount smoked. The authors stated, that despite the short follow-up and under-reported physician deaths, “we thought it necessary, in view of the nature of the results, to lay these preliminary observations before the survivors of the 40,000 men and women who made them possible.” [1]

The 50-year follow-up of these physicians confirmed the suspected relation of smoking to 12 of the 13 types of cancer previously thought related. [2]

This study by Doll and Hill has entered the realm of classics in epidemiology, as one of the icons of cohort studies (e.g. British Physicians Study; the Framingham Study; the Seven Countries Study).


[1] Doll R, Hill AB (1954). The mortality of doctors in relation to their smoking habits. British Medical Journal, 1 (4877), 1451-55.

[2] Doll R, Peto R, Boreham J, Sutherland I (2005). Mortality from cancer in relation to smoking: 50 years observations on British doctors. British Journal of Cancer, 92, 426-429.