Click on the links below to access all the ArticlePlus for this article.
Please note that ArticlePlus files may launch a viewer application outside of your web browser.
SD:Bearing in mind that few people start their careers with the intention of becoming epidemiologists, I wonder if you could describe how you came into the field yourself.
RD: I came into epidemiology through a series of fortunate accidents. At the end of my medical training I wanted to become a brain surgeon. However, the war interrupted my plans and, by the time the war was over, it was too late for me to start a long period of training in neurology and surgery. There were very few research jobs available in Britain immediately after the war and a friend of mine (who subsequently became my wife) introduced me to Dr. Avery Jones, a gastroenterologist at the Central Middlesex Hospital. Avery Jones was looking for an assistant in an epidemiologic study of the etiology of peptic ulcer, actually an occupational survey were looking at how the incidence of peptic ulcer varied with people’s hours of work—night work, day work, shift work, etc. 1 This seemed to me an excellent opportunity and so I took it.
I had always had an interest in mathematics and the possibility of applying mathematics to medicine and, while I was carrying out the peptic ulcer study, I had the opportunity of taking a course in medical statistics at the London School of Hygiene and Tropical Medicine. Here I met Professor Bradford Hill, who asked if I would go to work with him in a study he was about to set up to investigate the causes of lung cancer, after the peptic ulcer study was finished. So that is how I came into the field.
SD:How would you describe the subject of epidemiology as it was practiced when you started your career?
RD: Just after the war there really wasn’t a subject of epidemiology in Britain. However, quite a few young people were concerned about the social causes of disease. We were fortunate in that the then Secretary of the Medical Research Council was sympathetic to the idea that social and environmental factors might be important causes of disease, and he set up several units investigating such factors which, in effect, turned into Departments of Epidemiology. One of these was the Social Research Unit under Jerry Morris at a London hospital. Professor Ryle at Oxford was, however, really the first to make the term “social medicine” acceptable in university circles. Young people, such as Archie Cochrane, Jerry Morris, Donald Reid, John Pemberton, myself and several others, gathered around those few senior people who were interested in developing the subject. The whole world was open to us because there had been so little exploration of the field at that time.
SD:If you had to single out one person as having the strongest influence on your career, who would that be?
RD: There is no question that it would be Tony Bradford Hill. His actual name was Austin, but he was always known as Tony. He had always wanted to be a doctor, but developed tuberculosis in World War I and was unable to undertake the training, so went into medical statistics under the guidance of Major Greenwood, who had been a friend of his father’s. Tony did not have a degree in either medicine or statistics, but had an enormous interest in both subjects and became first Reader and subsequently Professor of Medical Statistics at the London School of Hygiene.
Tony had a genuine deep interest in medical problems and always sought to learn the relevant medicine whenever he undertook a new subject. He also set great store on data validity. He did not believe in sitting back and analyzing other people’s data and thought that it was the research worker’s own responsibility to have clean and reliable data. Tony didn’t like complex statistical techniques; he described statistics as the application of common sense and liked to be able to explain what was being done in his analyses of the data without algebra in such a way that a nonstatistical person could understand it. It was Tony’s emphasis on finding out the genuine medical importance of a problem, on the validity of the data that you were using and on simplicity of analysis that were really the hallmarks of his teaching.
SD:Collaboration is essential to successful epidemiologic studies. Personally, what have you found to be the most important ingredient of a good collaboration?
RD: I think the most important point is that everybody taking part should understand what the objective is, and should have an opportunity for influencing what they are doing. You can’t collaborate with people just by asking them to collect data or carry out analyses on your behalf. Collaboration means mutual trust and an understanding of what the contribution from each person is going to be.
I have been fortunate in the people that I have been able to collaborate with: for example, Michael Court Brown, the radiobiologist and a most unusual character; Avery Jones, the gastroenterologist; Bradford Hill; Malcolm Pike, who came to work with me as a statistician; Richard Peto, an outstanding colleague; and then, in the radiation field, Sarah Darby. With all these people, I have been able to work on a level basis where each fully understands what his/her own part is and what the other is doing. For more extensive collaborations involving many people in many hospitals or units, the same principles apply. The collaborators must be made to feel that they are essential members of the team and not just providers of data.
SD:I wonder whom you would regard as the most important epidemiologist during your lifetime?
RD: There is no doubt that this would again be Bradford Hill, as he laid down so many of its principles. The most important of these is the use of common sense rather than the mechanical application of formulae, and all his successors in Britain have tried to follow his example. I have already stressed the importance he gave to a thorough understanding of all the data that constitute your material and to a thorough understanding also of what you are doing with the numbers when you apply statistical techniques. But Bradford Hill also stressed that problems as a whole should be approached with common sense. Each problem has to be tackled first through achieving a thorough understanding of its nature and then by working out what is the most direct and simplest way of achieving an answer. The conclusion as to whether an observed association is causal or not is one of the most important and difficult things an epidemiologist has to do. And the logic of reaching that conclusion needs thinking through afresh every time in relation to a new problem.
SD:Turning now to your own work, which of your many studies would you say has been the most influential?
RD: This has to be the follow-up study of British Doctors, in which we obtained smoking histories from some 40,000 individuals, of whom 34,000 were male. We have followed this group since the early 1950s, with periodic new information about changes in their smoking habits. 2–15 In fact, I am now working on the 50-year follow-up. I think this study has been so influential because its results have contained so many surprises. It was originally set up to check the relationship between smoking and lung cancer that had previously been found by us 16,17 and others in case-control studies. Our own view was that the case-control studies produced sufficiently clear evidence by themselves, but many others considered this confirmatory evidence to be very important. However, the British Doctors’ Study didn’t just confirm the case-control studies—it also showed, surprisingly, that many other diseases might be caused by smoking, and the most recent follow-up has shown how prolonged cigarette smoking from early in life produces a much greater effect than one would have suspected from the early observations, where a cigarette smoker was often someone who had smoked only a pipe or cigars for the first 10 or 20 years. The British Doctors’ Study demonstrated very clearly that the regular smoking of cigarettes over a long period is very dangerous, much more so than the smoking of other tobacco products. Of course it is not the only study to have produced such information, and there have been enormous studies in America with much bigger numbers, which have shown the same thing. But certainly, I would say that, of the studies that I have had the privilege to take part in, the British Doctors’ Study has been the most important.
SD:At the other end of the spectrum, which do you think has been your most underappreciated paper?
RD: My favorite paper is the one with Michael Court Brown deriving the dose-response relationship between radiation and leukemia. 18 It wouldn’t be fair to say it has not been appreciated because both the Medical Research Council, which asked us to carry out the study, and a number of other people were very interested in our results and took them seriously. In addition, it was one of the studies that were mentioned by the United Nations when I was awarded the United Nations Award for Cancer Research. But this work is not often remembered nowadays. I doubt if many epidemiologists would pick the dose-response relationship between radiation and leukemia if they were asked to associate particular observations with my work. Yet, it is certainly the second most important piece of work that I have done, after the effects of smoking, and it provided the first suggestive evidence of a linear relationship for the carcinogenic effect of exposure to ionizing radiation down to quite small doses. In fact, the estimate of the risk of leukemia per unit dose that it provided is not very different from the value that is accepted now.
It was quite a difficult study to organize. We had to collect information on 14,000 patients treated throughout the country by radiotherapy for the benign condition of ankylosing spondylitis, and we had to measure the dose of radiation received in the marrow by doing experiments on a model man. In many ways it was the best-designed study I have ever participated in and possibly my best work.
SD:Your paper with Michael Court Brown on the relationship between radiation and leukemia was published as a Medical Research Council report. Was it also published in a journal?
RD: No, it was never published in a journal. All the important information was included in the Medical Research Council report.
SD:Do you think that this might have played a role in peoples’ not appreciating it? Presumably it would be quite hard to get a hold of that report these days, especially for people outside Britain.
RD: Yes, probably, although later extensions of the study were published in ordinary scientific journals. 19–29
SD:Which of your contributions to epidemiology would you most like to be remembered for?
RD: Well, I would have to say my work on the effect of smoking, because it has been of such practical importance; but I should like to think that I was also remembered for some other things, such as my work on oral contraceptives and asbestos.
SD:You have clearly spent a large proportion of your time working on epidemiology. What have been your major interests apart from this?
RD: Well, I suppose my home has been my major interest and source of enjoyment. I have also been interested in literature and the theater, and I enjoy traveling. I did, however, get involved in establishing a new college in Oxford, called Green College after its principal benefactor, and helping it to develop has been a very major interest.
SD:What would you say are the most essential elements of an epidemiologist’s working environment for a productive career?
RD: The possibility of collaboration is certainly vital. Collaboration with people who have a good understanding of the major medical problems in the world is, I think, the most important thing for an epidemiologist. Of course, it is becoming increasingly difficult to solve problems of etiology by epidemiologic methods as the easy things have been done. I am certain there still remain many problems that can be solved, but it is necessary to collaborate with specialists in the relevant fields to solve them.
Nowadays, of course, access to good computing facilities and knowledge of how to use them is also very important, whereas computers were nonexistent when I entered the subject.
SD:In your opinion, what has been epidemiology’s most important contribution to society as a whole?
RD: I don’t think there has ever been anything as important as Snow’s discovery that fecal contamination of water causes cholera. This led to the discovery of all waterborne diseases and provided the possibility of preventing an enormous amount of ill health throughout the world. I don’t think there is anything else comparable with it, and some of the lessons from it have yet to be put into effect.
SD:What would your assessment be of the current state of the health of epidemiology?
RD: In Britain, it’s not very good. The criteria that are laid down now for advancement from junior to assistant to professor are very difficult for an epidemiologist to progress through in this country. This is because of the way that salaries are determined and the need, if you are medically qualified, to get classified as a consultant. There are also so many bureaucratic obstacles now to carrying out epidemiological studies per se. I was very fortunate to be able to work without having to worry about them.
Universities, too, have become so enthralled with molecular genetics that senior people tend to relegate epidemiology to a very minor corner of medical schools. I think this is an error because molecular genetics is still going to require epidemiological observations to determine the important causes of human disease.
SD:What would you say are the biggest risks that our profession faces at the moment?
RD: To my mind, one of the biggest risks is that of becoming overstatistical and of thinking that everything can be solved by means of computers. I have seen, as I am sure all epidemiologists have, young people applying complex statistical formulae without really understanding what is going on in the modeling behind the results that they end up with. I think this remains one of the greatest hazards in epidemiology at the present moment.
SD:Would you see the increasing difficulty in getting access to data through regulations related to confidentiality and data protection as a risk, as well?
RD: Indeed, yes. In fact, I should have mentioned that first when you asked about the risks that our profession faces. There has been an enormous change in the attitude towards confidentiality. I wouldn’t say that it was a change in the public attitude, so much as in the governmental attitude, because I am not sure that the public is really as concerned as governments appear to be. When I started in epidemiology, we operated on the old system, which was approved by the Medical Research Council, that a doctor could pass information about a patient to another doctor, relying on the fact that he or she would be bound by the Hippocratic oath to treat details about patients confidentially. We had no difficulty in collecting all sorts of information as long as the process was covered by someone medically qualified. I say “covered by” because quite often the statistician would be the one actually handling the data but, in order to meet the conditions at the time, you had to have a medical person accepting the responsibility for the confidentiality of the data. This system worked perfectly well, and I know of no trouble having been caused for anybody by the free passage of information between clinicians and epidemiologists for the study of disease.
Now it is becoming horrifyingly difficult to get hold of epidemiological information relating to individuals, and I can see great difficulties for the epidemiologists of the future. I didn’t immediately mention this point when you first asked about risks faced by our profession because I find it so depressing that I have suppressed it in my mind. It won’t affect me, as I shan’t be involved in research in 10 years’ time, but valuable research of importance to the public health really is being made extremely difficult, if not impossible in some cases.
SD:On a more optimistic note, what would you see as the ripest opportunities in epidemiology at the moment?
RD: I don’t think that at my age I should be speculating about this.
SD:Do you have any predictions about what the future may hold for our field?
RD: I have given up predicting the future for many years. I have nearly always been wrong, except in the case of the necessity of war with Hitler’s Germany in the 1930s, when young people such as myself saw the horrors of what was going on in Germany and realized there was no alternative. Most of my other predictions have turned out to be incorrect.
SD:As a final question, I would like to ask what would be the most important piece of advice you would give to a young person starting out on a career in epidemiology?
RD: I have no doubt that the most important thing for a young person to do is to get attached to a good epidemiological unit. Preferably, the best one there is in the country. He or she is going to learn by seeing how epidemiology is practiced by colleagues, and the best way to do this is by working with the best epidemiologists. So what I would say to a young epidemiologist is go for the top and work with the best people.
The journal gratefully acknowledges the help of David Harwood and Oxford Medical Illustration of the John Radcliffe Hospital in making possible the video recording of this interview. We also thank Nina Keleher for secretarial assistance.
About the Interviewer
SARAH DARBY is a statistical epidemiologist who works in the Clinical Trial Service Unit of the University of Oxford. She first worked with Richard Doll over 20 years ago on extending the follow-up of the patients treated with radiotherapy for ankylosing spondylitis. Since then she has worked with him on a number of other radiation-related topics, including studies of the U.K. atmospheric nuclear weapons tests and studies to estimate the carcinogenic effect of residential radon.
1. Doll R, Jones FA, Buckatzsch MM. Occupational factors in the aetiology of gastric and duodenal ulcers. Med Res Council Spec Rep Ser No. 276. London: HMSO, 1951.
2. Doll R, Hill AB. The mortality of doctors in relation to their smoking habits. BMJ 1954; 1: 1451.
3. Doll R, Hill AB. Deaths from poliomyelitis among British doctors. BMJ 1957; 1: 372.
4. Doll R, Hill AB. Mortality in relation to smoking: 10 years’ observations of British doctors. BMJ 1964;1:1399–1410 and 1460–1467.
5. Doll R, Hill AB. Mortality of British doctors in relation to smoking: observations on coronary thrombosis. In:Study of Cancer and Other Chronic Diseases. National Cancer Institute Monograph No. 19. Bethesda, MD: National Cancer Institute, 1966.
6. Fletcher C, Doll R. A survey of doctors’ attitudes to smoking. Brit J Prev Soc Med 1969; 23: 145–153.
7. Doll R, Pike MC. Trends in mortality among British doctors in relation to their smoking habits. J Roy Coll Phys London 1972; 6: 216–222.
8. Doll R, Peto R. Mortality in relation to smoking: 20 years’ observations on male British doctors. BMJ 1976; 2: 1525–1536.
9. Doll R, Peto R. Mortality among doctors in different occupations. BMJ 1977; 1: 1433–1436.
10. Doll R, Peto R. Cigarette smoking and bronchial carcinoma: dose and time relationships among regular smokers and lifelong non-smokers. J Epidem Comm Hlth 1978; 32: 303–313.
11. Doll R, Gray R, Hafner B, Peto R. Mortality in relation to smoking: 22 years’ observation on female British doctors. BMJ 1980; 280: 967–971.
12. Doll R, Peto R, Wheatley K, Gray R, Sutherland I. Mortality in relation to smoking: 40 years’ observations on male British doctors. BMJ 1994; 309: 901–911.
13. Doll R, Peto R, Hall E, Wheatley K, Gray R. Mortality in relation to consumption of alcohol: 13 years’ observations on male British doctors. BMJ 1994; 309: 911–918.
14. Doll R, Peto R, Hall E, Wheatley K, Gray R. Alcohol and coronary heart disease reduction among British doctors: confounding or causality? Eur Heart J 1997; 18: 23–25.
15. Doll R, Peto R, Boreham J, Sutherland I. Smoking and dementia in male British doctors: prospective study. BMJ 2000; 320: 1097–1102.
16. Doll R, Hill AB. Smoking and carcinoma of the lung. BMJ 1950; 2: 739.
17. Doll R, Hill AB. A study of aetiology of carcinoma of the lung. BMJ 1952; 2: 1271.
18. Court Brown WM, Doll R. Leukaemia and aplastic anaemia in patients irradiated for ankylosing spondylitis. Med Res Council Spec Rep Ser No. 295 London: HMSO, 1957.
19. Court Brown WM, Doll R. Mortality from cancer and other causes after radiotherapy for ankylosing spondylitis. BMJ 1965; 2: 1327–1332.
20. Court Brown WM, Doll R, Smith P. Neoplasia in patients treated with X-rays for ankylosing spondylitis or metropathia haemorrhagica. In: Harris RJC, ed. Proceedings of the IXth International Cancer Congress. UICC Monograph Series, Vol. 10. Berlin: Springer-Verlag, 1967;119–126.
21. Smith PG, Doll R, Radford EP. Cancer mortality among patients with ankylosing spondylitis not given x-ray therapy. Brit J Radiol 1977; 50: 728–734.
22. Radford EP, Doll R, Smith PG. Mortality among patients with ankylosing spondylitis not given x-ray therapy. N Engl J Med 1977; 297: 572–576.
23. Smith PG, Doll R. Age and time dependent changes in the rates of radiation-induced cancers in patients with ankylosing spondylitis following a single course of X-ray treatment. In: Late Biological Effects of Ionizing Radiation, Vol. 1. Vienna: International Atomic Energy Agency, 1978.
24. Smith PG, Doll R. Mortality among patients with ankylosing spondylitis after a single treatment course with x-rays. BMJ 1982; 284: 449–460.
25. Darby SC, Doll R, Gill SK, Smith PG. Long term mortality after a single treatment course with x-rays in patients treated for ankylosing spondylitis. Brit J Cancer 1987; 55: 179–190.
26. Darby SC, Doll R, Smith PG. Trends in long-term mortality in ankylosing spondylitics treated with a single course of X-rays. Health effects of low-dose ionising radiation—recent advances and their implications. In:Proceedings of the International Conference, London, 11–14 May 1987. London: British Nuclear Energy Society, 1988;51–56.
27. Lewis CA, Smith RE, Stratton IM, Darby SC, Doll R. Estimated radiation doses to different organs among patients treated for ankylosing spondylitis with a single course of x-rays. Brit J Radiol 1988; 61: 212–220.
28. Weiss HA, Darby SC, Doll R. Cancer mortality following x-ray treatment for ankylosing spondylitis. Int J Cancer 1994; 59: 327–338.
29. Weiss HA, Darby SC, Fearn T, Doll R. Leukaemia mortality following x-ray treatment for ankylosing spondylitis. Radiation Res 1995: 142; 1–11.
© 2003 Lippincott Williams & Wilkins, Inc.