Share this article on:

A Conversation With Brian MacMahon

Willett, Walter

doi: 10.1097/01.ede.0000129522.22687.1b

Supplemental Digital Content is Available in the Text.

Editor's Note: VOICES is a project of the journal to provide personal, historical and scientific perspectives on the field of epidemiology, as seen through the eyes of the field's most senior and accomplished practitioners. Subjects are selected by the Editors. Readers are welcome to nominate candidates for Voices.

Brian MacMahon's curriculum vitae is available with the online version of the Journal at

This interview was conducted 8 November 2003, at Dr. MacMahon's home in Needham, Massachusetts. Brian MacMahon has approved this transcript for publication.

Correspondence: Walter Willett, Department of Nutrition, Harvard School of Public Health, 655 Huntington Ave., Boston, MA 02115.

Back to Top | Article Outline


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.

Brian MacMahon was born on August 23, 1923, in Sheffield, England. At the University of Birmingham, England, he earned the diplomas of the Royal Colleges of Physicians and Surgeons in 1946, the MB, ChB degree in 1948 and the MD degree in 1955. Between 1946 and 1948, MacMahon served as Ship Surgeon in the English Merchant Navy. He obtained a PhD in Social Medicine at Birmingham, UK, and then a Master's degree in Epidemiology at Harvard School of Public Health in 1953. After faculty appointments back in Birmingham and in Brooklyn, NY, he was appointed Head of the Department of Epidemiology at Harvard School of Public Health in 1958. In 1960, with his colleague Tom Pugh, MacMahon published Epidemiologic Methods, later to become Epidemiology: Principles and Methods,1 which was the first modern epidemiology textbook and served as the standard text for decades. At Harvard, MacMahon built a department that developed national and international leadership in epidemiologic methods, particularly with respect to cancer. His graduates have become leaders in academia and governmental research units worldwide. He is most widely known for his research on the etiology of breast cancer, which included a novel international case-control study. In 1992, Dr. MacMahon received the Charles S. Mott Prize from the General Motors Cancer Research Foundation for his research on cancer causation and prevention. He has received honorary doctorate degrees from the University of Athens, Greece, the State University of New York, and the University of Birmingham, England. Dr. MacMahon retired in 1988, and has his home in Needham, Massachusetts.

Back to Top | Article Outline


WW: Few people start with the intent of becoming an epidemiologist. Can you describe your own path that brought you to epidemiology?

BM: After a marginally satisfactory performance in medical school, some evening locum work in one of the poorer districts of Birmingham, and a couple of years as a ship's doctor, I became somewhat dispirited about a future in clinical medicine. I enrolled in the DPH program (U.S. equivalent MPH) at the University of Birmingham with the intention of going into public health. In that program and in a subsequent PhD program in the same department, I came under the influence of a rather remarkable trio consisting of Tom McKeown, Ron Lowe, and Reginald Record. Most influential on me were McKeown and Lowe. They proposed (and, together with Record, supervised me in) a study of infantile pyloric stenosis2—a clinically dramatic disease of early infants that had some very interesting demographic features. It was much more common in males than females, was more common in first births, and appeared to run in families. That was my first contact with what I subsequently came to think of as epidemiology, but was called at that time, at least in Britain, social medicine.

Figure. B

Figure. B

I continued an interest in this line of work—although now directed toward leukemia and breast cancer—with Duncan Clark in the Department of Environmental Medicine and Community Health in the State University of New York, Downstate Medical Center in Brooklyn. (Duncan had borne principal responsibility for my migration to this country.) I became a card-carrying epidemiologist only when invited in 1958 to be chairman of the department of that name at the Harvard School of Public Health. My research interests remained essentially unchanged, but they were now legally epidemiology. I have been content in that designation.

WW: What is the biggest difference between epidemiology as it was practiced when you started your career and epidemiology today?

BM: In the matter of data collection and analysis. I should tell you that in my PhD program, I conducted interviews personally with approximately 500 mothers of infants who had had the disease and a comparable group of controls. I conducted these interviews in the mothers’ homes and transferred the responses to blank 3 × 5 cards, which were then sorted into groups and counted—by me! Marginal “punch cards,” particularly the “advanced” McBee cards (with 2 holes punched for each field of information), came later. I maybe had a mechanical calculator for the analysis, but at that time, to have any kind of calculator, mechanical or electric, on your desk carried the same kind of cachet that having a computer has today. It was very much a pencil-and-paper operation.

Today, of course, there are few epidemiologists who collect their own data. Generally, epidemiologists have a great deal to say about what data are collected and by whom, but usually the data are collected and processed by other professionals, sometimes nonprofessionals. They may be analyzed and interpreted by the epidemiologist who may use very sophisticated computing equipment.

The subject matter emphasis of epidemiology has also changed considerably. There is still very active work on the epidemiology of the infectious diseases, but the emphasis on these areas has shifted from departments of epidemiology to departments of microbiology and others, with epidemiology turning its attention largely to the noninfectious diseases (or as my colleague, Tom Pugh, called them, “diseases not yet known to be associated with infectious agents”).

More broadly, the use of what is now called “epidemiologic” evidence is given much greater recognition both in medicine and in society generally. It is routinely called on in legal and regulatory proceedings, which certainly was not the case when I entered the field. Sometimes even the press gives us space—not always to our advantage.

WW: Who would you single out as having most strongly influenced your career?

BM: No single person. I would have to start with my father, who pushed me into medical school against my better judgment. I was interested in music and history, but in that day and age you did what your father told you and so I went to medical school. In graduate work after medical school, I was greatly influenced by the people I have already mentioned, Tom McKeown and Ron Lowe, and it was there that, with Record's help, I think I learned, as the carpenters say, “To measure twice and cut once.” To put it in epidemiologic terms, think n times and come to a conclusion only once—and then only for a short time until the next body of data becomes available.

WW: Collaboration is essential to successful epidemiologic studies. What have you found to be the most important ingredients to good collaboration? With whom have you had your best collaborations?

BM: That is 3 questions. I do not think I can remember them. Do not forget that I am 80 years old. My memory, particularly, is not functioning as it used to.

WW: What have you found to be the most important ingredients for good collaboration?

BM: The first question was?

WW: That was just a statement that collaboration is essential….

BM: I do not entirely agree with that statement. I think collaboration is important and tremendously beneficial to epidemiology. I do not think it is essential. A lot of good epidemiology has been done as individuals, at least in the past (I need mention only John Snow). It may be much more true today that collaboration is essential.

Figure. B

Figure. B

WW: What have you found to be the most important ingredients of a good collaboration?

BM: Nothing more than the things that are important in any personal relationship. Trust, openness, genuine participation. Mutual respect helps.

WW: With whom have you had your best collaborations?

BM: That is hard to say. The best collaborations I have had have been with my own colleagues in the same department. Outside of that, I think the most fruitful collaboration was with the group that we set up to study breast cancer sometime in the early 1960s.

WW: Would you just want to mention some of the people and where they were? Some of them did become your permanent colleagues.

BM: They are really too many to remember, but among them I recall Phil Cole, Tomio Hirohata, Bob Hoover, Tom Mack, Ken Rothman, Dimitrios Trichopoulos, Noel Weiss, and, of course, yourself.

WW: Whom would you regard as the 2 or 3 most important epidemiologists during your lifetime?

BM: Without question during the latter half of the 20th century Richard Doll. He was the preeminent epidemiologist in terms of both his own research and his influence on the field. Doll himself recognized his debt to Bradford Hill, but I had no personal contact with Bradford Hill. After Doll, I would not like to pick the second or third from among the many who have played influential roles.

WW: Which do you think has been your most influential paper?

BM: The paper on the age of first birth and breast cancer risk,3 which I think stimulated research on possible hormonal determinants. Also, the series of papers on pyloric stenosis of infants4 helped shift research on etiology from genetic to environmental factors.

WW: Which has been your most underappreciated paper?

BM: I have about 200 of those, and I do not know which to pick. One paper that I have been sorry not to see further developed is that on the bimodal age distribution of Hodgkin disease.4 This paper provided evidence that what was then (and still is) considered one disease contains at least 2, and probably 3, major etiologic entities. One of the major component diseases, which I speculated was associated with a microbiologic agent, occurred predominantly in young adults. Nancy Gutensohn (Mueller) at Harvard has followed that up and expanded the idea with biologic studies, but I am not aware of anybody else who has acted on the implication of that paper.

WW: How have you chosen to research questions you want to address?

BM: There is a very fine decision involved in balancing what you think is an important problem in numerical terms with the belief that you may make a contribution to some measure of action against the problem. I think that is why I chose breast cancer—it combines a disease that clearly is of numerical importance with one that has striking demographic features that need to be explained. My ambition to explain those features remains unfulfilled.

WW: Which of your contributions would you most like to be remembered for?

BM: No specific contribution of my own, but the fact that I was able to provide a haven for a succession of promising investigators, many of whom are now major figures in the field of epidemiology. In this I had the support of my institution, particularly Jack Snyder and Jim Whittenberger who were Deans during my early days, and the National Institutes of Health and their financial resources.

WW: Would your book 1 be another contribution? That book, in some sense, codified the whole field of chronic disease epidemiology for decades.

BM: Yes—immodestly, I agree with that characterization. Certainly, it was more important than any of my personal characteristics in leading productive people in the direction of epidemiology.

WW: What have been your major interests outside of your epidemiologic activities?

BM: Well, my home and my children have been my major interest and my major source of pleasure. I did a lot of camping and hiking with them in earlier years. That has remained very much with me, in my mind if not in my physical body. Outside of that, music has been my main preoccupation.

WW: You play the piano yourself.

BM: I did, at one time, play the piano and also the clarinet. I gave up the clarinet many years ago. I persisted with the piano until approximately 10 years ago when I developed Dupuytren's contracture in both hands. With that, it is not possible to play the piano and difficult even to play the computer.

WW: In your opinion, what is epidemiology's most important contribution to society?

BM: In recent times, I believe that it is the discovery of the major preventable feature of lung cancer and other diseases related to cigarette smoking. However, it depends on how far you want to go back. Snow's work on cholera in London was, of course, tremendously influential in control not only of cholera, but of a lot of waterborne diseases. The elegance of Snow's writing was also very important to the development of epidemiology itself. One wonders if people would have the motivation to study lung cancer if they had not had the examples of Snow, of Semmelweis on puerperal fever, and other earlier pioneers to look back on.

WW: Not just the specific discoveries, but the concept that causes of disease can be identified and changed?

BM: Yes, with the important proviso that diseases as they now exist (and we see this with lung cancer) have multiple causes. Cigarette smoking is a major cause, in the sense that the disease would not happen so frequently if smoking did not exist. That is the sense in which I use the word cause.

WW: There is still a contribution there that you are pointing to, a broad concept that epidemiologists have provided, regarding multiple causality or multiple causes.

BM: I am not sure how widespread that concept is even in epidemiology, but certainly it should be. What we should ultimately be searching for are experiences that can be prevented or augmented in such a way as to change the frequency of disease—whether they increase it, as when we look at toxins such as asbestos, or whether they decrease disease, as we usually think of today when we look for preventive causes in diets. Not that the removal or augmentation of one specific cause will necessarily change the frequency of the totality of disease as it is currently defined, but if it just produces a change in the frequency in that disease, we have accomplished something.

WW: What is your assessment of the current state of health of epidemiology?

BM: Richard Doll has expressed some pessimism about this in Britain (which seemed largely to result from administrative problems). I am bullish about the future of epidemiology in the United States. It is my impression that the increase in financial support for epidemiology in recent decades has been even greater than that for most medical sciences, although recently this leading role has been taken over by genetics. Even so, support for epidemiology is still strong, and there is evidence in my mind that it will continue. I think there is increasing acceptance of the potential contribution of epidemiology in the medical profession, in grant-giving agencies, and in the public debate.

WW: What are the biggest risks our profession faces?

BM: The biggest potential risk is the current, in my mind exaggerated, concern for privacy of medical and other records. I do not think it has been a serious problem so far, but it could well be if it continues to go in the direction it is going. This jeopardizes the whole possibility of conducting epidemiologic studies in the future. I do not know where this concern arises. To some extent, of course, the concern is reasonable, but when it threatens societally beneficial efforts, it is not. I have a strong feeling that the strident rhetoric that we sometimes hear does not reflect the feelings of the general population. We ourselves conducted many, many thousands of interviews (I have never estimated the total, but certainly it is more than 100,000) with patients and randomly selected people in the Boston area and in several other areas abroad. I know of no instance in which there was harm done to any who were interviewed. Of course, a few, generally on the order of approximately 10% of those approached, declined to be interviewed or provide information. That is their right, and I have no difficulty with it. The great majority of people are not as possessive about the information on their past illnesses or even their relatives’ past illnesses as whoever it is who is promoting this ultimate-privacy movement seem to think. It seems to me to be a push by a very small group, probably with political connections of some kind. I do not know what their motivations are, but they are a present, and potentially even greater, problem for epidemiology.

WW: What are the ripest opportunities for epidemiology?

BM: I think we mostly covered that already, in a general sense. Opportunities for collaboration not only with people in the clinical fields, but also with those people in basic sciences. I think they are expanding. To the extent that those kinds of collaboration can be developed, I think they provide strong opportunities for epidemiology.

WW: What would be the single most important piece of advice you could give to new epidemiologists starting their career?

BM: I cannot do better than echo what Doll told his interviewer in this series. To find—he said, “a group,” but I would say either a person or a group—that is experienced and productive. He did not mention productive. I think that is important also. Attach yourself there as an observant apprentice.

Back to Top | Article Outline


Walter Willett is Professor of Epidemiology and Nutrition at Harvard School of Public Health and Chair of the Department of Nutrition. He worked with Brian MacMahon as a doctoral student over 25 years ago, which stimulated his interest in the potential relation between dietary factors and cancer incidence. Following this lead, Willett developed dietary assessment methods that are being used in several large cohort studies to investigate the relation between nutritional factors and most major health outcomes.

Back to Top | Article Outline


Thanks to Judy Eshelman who transcribed the interview and to Debbie Flynn and Hilary Farmer who prepared the edited manuscript and the accompanying materials.

Back to Top | Article Outline


1. MacMahon B, Pugh TF. Epidemiology: Principles and Methods, 1st ed. Boston: Little, Brown and Co; 1970.
2. MacMahon B, Record R, McKeown T. Congenital pyloric stenosis: an investigation of 578 cases. Br J Soc Med. 1951;5:185–192.
3. MacMahon B, Cole P, Lin TM, et al. Age at first birth and breast cancer risk. Bull. World Health Organ. 1970;43:209–221.
4. McKeown T, MacMahon B. Infantile pyloric stenosis in parent and child. Arch Dis Child. 1955;30:497–500.
5. MacMahon B. Epidemiological evidence on the nature of Hodgkin's disease. Cancer. 1957;10:1045–1054.

Supplemental Digital Content

Back to Top | Article Outline
© 2004 Lippincott Williams & Wilkins, Inc.