VIEWPOINT

Viewpoint

Should the mission of epidemiology include the eradication of poverty? Kenneth J Rothman, Hans-Olov Adami, Dimitrios Trichopoulos Physicists seem to have escaped the old criticism that their work is impractical. Perhaps the criticism was blunted by technological innovations that rest on physical theory. Nevertheless, even astrophysicists, whose work seldom induces engineering breakthroughs, can now pursue knowledge for its own sake without fear of being badgered about the practical relevance of their work. What physicists have gained, however, epidemiologists seem to have lost. Accusations have been mounting that epidemiologists have abandoned their public-health mission of being “physician-scientist” to society1 in favour of studying the scientific arcana of disease causation. Editors at The Lancet suggested last year that the “woolly breadth” of this criticism should be replaced by a detailed proposal for what should be done to put “public health back into epidemiology”.2 We believe that, contrary to what the critics suggest, epidemiology has been evolving in line with its public-health mission, and that the criticism itself—however noble in its motivation—is not coherent. The critics allege that epidemiologists have lost their traditional social perspective; instead, they argue, epidemiologists have become mired in statistical procedures and molecular mechanisms in the study of individual risk factors of lesser and lesser importance. According to Shy, who deigned to put “academic epidemiology” on trial, “the focus of [epidemiology] still remains on understanding the proximate and individual rather than the underlying and societal determinants of disease”.3

Accusation 1: epidemiology is too individualistic As with other biomedical sciences, epidemiology yields practical knowledge. Many applications can be carried out directly by individuals. For example, information about the risks of having unprotected sex can persuade people to change their sexual behaviour, and information about the risks of smoking can motivate smokers to give up their habit. These actions are, to some extent at least, personal choices, which public-health programmes can influence through educational materials, as they have done through campaigns on smoking and health. Other epidemiological knowledge, however, cannot be easily applied without actions at societal level. Thus,

smallpox could not have been eradicated without a clever, global strategy to contain it, and malnutrition rooted in poverty cannot be prevented without societal interventions that ease the burden of poverty or that address malnutrition directly. The distinction between individual and societal applications of epidemiological knowledge are at the core of the new wave of criticism. The central complaint is that epidemiologists have focused on individual risk factors to the exclusion of broader societal causes of disease. Thus, wrote Shy, epidemiologists “tell us that a lower fat content of the diet, a lower population-average body weight, or a lower prevalence of smoking will reduce the population risk of coronary disease. But they do not explore societal policies, incentives, or other interventions at the ‘organised community’ level that would bring about a shift in the population distribution of fat consumption, body weight, or smoking habits”.3

Pearce, one of the harshest critics of epidemiologists, portrays this slant as a personal and political choice.4 In his view, it is not so much the lure of science as an end in itself that has swivelled epidemiologists to a biomedical orientation; rather it is a tide of political conservatism and a personal indifference to the health problems of others that account for the new direction in epidemiology. According to Pearce, epidemiologists are so self-indulgent that they prefer to study “decontextualised individual risk factors” instead of “upstream” causes of health problems, such as poverty, because “epidemiologists tend to be most interested in risk factors that they can relate to, or may even be exposed to. Epidemiologists are frequently at risk from tobacco smoke, alcohol, diet, viruses, and even some occupational chemical exposures, but they are rarely at risk of being poor”.4

Pearce has gone so far as to criticise the dogged reliance of the biomedical establishment on randomised, controlled trials. Whereas many scientists believe that a randomised trial is a crucial tool in helping distinguish causes from confounding factors, Pearce claims, without any illustration, that randomised trials can be dangerous instruments of “decontextualisation” that are “inappropriate in studies that require a consideration of the historical and social context. The danger is that attempting to eliminate the influence of other causes of diseases—in an attempt to control confounding—strips away the essential historical and social context . . .”.4

Lancet 1998; 352: 810–13 Department of Medicine and Epidemiology, Boston University Medical Center, Boston, MA, USA (Prof K J Rothman DrPH), Department of Epidemiology, Harvard School of Public Health, Boston, MA, USA (K J Rothman, Prof H-O Adami MD, Prof D Trichopoulos MD); and Department of Medical Epidemiology, Karolinskay, Institute, Stockholm Sweden (H-O Adami) Correspondence to: Prof Kenneth J Rothman, Boston University, School of Medicine, Evans Department of Medicine, Section of Preventive Medicine and Epidemiology, Boston, MA, 02118-2526, USA

810

Accusation 2: epidemiologists need more moral and political fibre Critics contend that epidemiologists have a greater responsibility than merely to study the causal role in human health of factors such as poverty or tobacco consumption. In their view, epidemiologists must also strive to eradicate the upstream causes of health problems. For example, if epidemiological research indicates that poverty causes malnutrition, which in turn causes infant

THE LANCET • Vol 352 • September 5, 1998

VIEWPOINT

death, the epidemiologist’s responsibility is to work towards the elimination of poverty. This activity requires the lumbering apparatus of social and political forces to be set in motion—something that most epidemiologists have been loath to attempt. Critics claim that today’s epidemiologists lack the moral resolve and political fire to complete their professional mission. In a 1997 book, Public Health at the Crossroads, Beaglehole and Bonita write5 “Unfortunately, for many epidemiologists the study of social factors is considered too political . . . it is necessary for epidemiology to affirm its connection with policy and to reject scientific isolation”.

Susser and Susser6 suggest that the perceived drifting of epidemiologists away from social commitment stems from a deficiency in their training. They argue “a practical programme must be devised to ensure that, in the course of their education, epidemiologists are socialised in a manner that keeps alive the idea of improving the public health as a primary value”.6

In short, some critics believe that epidemiologists of today lack a firm commitment to public health. Instead, claim the critics, they fritter away their professional time by studying scientific minutiae at the expense of urgent public-health problems.

pathogenicity of the Vibrio organism depends on the El Niño Southern Oscillation, and how cholera may spread between continents through ocean currents.9,10 For cholera, as for other diseases, the more knowledge we acquire of causal pathways at all points—from the most “fundamental” or “ultimate” social, political, and economic determinants to the molecular and biochemical determinants most proximal to disease occurrence—the better the foundation we lay for any effective public-health action. To observe, for example, that in regions where poverty is rife, a given disease occurs at a greater rate than in areas not affected by poverty, is surely not enough. Epidemiologists have been preoccupied with methodology mainly because they understand that such comparisons are often affected by innumerable biases that can lead to false inferences. Their preoccupation with methods is the inevitable evolutionary consequence of their drive to understand the causes of disease in the natural human environment—rather than in the laboratory cage or the petri dish. The same kind of concern that makes epidemiologists wary of ecological comparisons impels them to carry out randomised trials when experiments are ethical and feasible.

When is intervention most effective? Knowledge before action, or “Ready, Fire, Aim”? There is no denying that epidemiologists have progressively concentrated on the details of causal mechanisms. The only surprise to us is that anyone would regard this preoccupation with causal mechanisms as a problem. In past decades, epidemiologists could be criticised for studying mostly superficial relations between exposure and disease occurrence. Now the field is maturing, along with other biological sciences, and such superficiality is gradually being replaced with clearer insights into causal pathways. Whereas epidemiologists once studied factors such as dietary-fat consumption and total serum cholesterol, they have now progressed to classifying dietary fat by chemical structure—by comparison of low to high lipoprotein ratios—and are moving on to assessing the protective effect of antioxidants on fat-induced endothelial impairment.7 Is the preoccupation with causal mechanisms as detrimental as critics allege?—Not if reasonable knowledge of causation is deemed a sensible antecedent to intervention. If the moral purpose of epidemiology is to alleviate the human burden of disease, the primary task of epidemiologists should be to acquire insight into the causal chain, starting from root causes and continuing up through the beginnings of the disease itself. True, we do not necessarily need to know every detail about the pathway between intervention point and outcome; we can infer the effects of interventions even with gaps in understanding, just as John Snow showed how to curtail cholera before Vibrio cholera was identified. Knowledge is never perfect, and action is often indicated in the face of substantial uncertainty. Public-health professionals, however, do not have a license to tinker promiscuously with society. Public-health programmes may be conceived and implemented with great hope and yet turn out to be useless,8 or even damaging. One and a half centuries after Snow’s work on cholera, the public-health threat from the disease lingers, as research continues into how

THE LANCET • Vol 352 • September 5, 1998

Generally, the further upstream we move from the occurence of disease towards root causes, the less secure our inferences about the causal path to disease become. Even if our inference is correct, moreover, intervention with respect to upstream causes may be less efficient and therefore less effective than intervention closer to disease occurrence. Consider the causal path: poverty to malnutrition to infection to death. Control of infection at the level of the human individual may be the most efficient way to prevent death in this causal chain. True, if control of infection does not address the malnutrition underlying the infection, the infection is likely to recur; one might therefore reasonably look upstream to address the malnutrition problem. Nevertheless, if we attempt to combat malnutrition without also dealing with the concomitant infections, many people will die needlessly. Furthermore, there may be little that an epidemiologist can do in the short term to overcome malnutrition as a population problem. To awaken the public or political authorities to the problem of malnutrition and to redirect societal resources may eventually do the trick, but publichealth accomplishments mediated through social changes are won slowly. In the meantime, through their study of risk factors at the individual level and by the use of randomised trials, epidemiologists have discovered that vitamin-A supplements can prevent serious morbidity and many deaths in malnourished children.11–12 This knowledge will save lives, despite the fact that it does not alleviate gross malnutrition, does not require manipulation of any upstream cause, and is obtained from the type of epidemiological work that critics of epidemiology disparage.

What is the public-health solution to the poverty problem? The ultimate step in the preceding intervention scenario is to eliminate the poverty that causes the malnutrition. The critics urge that we take this step. It requires deep societal involvement in a laudable public-health end—one that any 811

VIEWPOINT

humane person would embrace. Yet, it is only fair to ask whether epidemiologists have the means to eradicate poverty. Is poverty eradication a public-health programme? How exactly should it be accomplished? Economists would seem the most likely candidates to supply the answer, which might go something like—“Let markets be free to expand without guiding them too firmly”. But other economists might give a different answer. Do the critics of epidemiology suggest that epidemiologists should lobby against international trade barriers, or in favour of them, in the pursuit of their public-health objectives? Perhaps the critics believe that epidemiologists should second-guess economists and attempt to eradicate poverty using their own epidemiological model. Given the scope of the task, sympathy might go to the epidemiologists who prefer to focus on a comparatively simple problem, such as the causes of cancer.

Did epidemiologists cause the global tobacco problem? Beaglehole and Bonita,5 and Pearce,4 cite the global tobacco problem as another example of epidemiological malfeasance. Their theme is that epidemiologists fiddle while the epidemic burns: epidemiologists have, ostrichlike, studied the effects of tobacco smoking well beyond the need to characterise this health menace. “[I]t can be argued that the fundamental problem of tobacco”, says Pearce, “lies in its production, rather than in its consumption”.4 Beaglehole and Bonita go further, blaming the worldwide tobacco problem in part on epidemiologists themselves, whose studies “focussed on consumption and not on production, have only added to the inequalities in health between the poor and rich within wealthy countries, and between poor and wealthy countries”.5 Apparently, they believe that epidemiologists would have done better to lobby against the planting and manufacturing of tobacco than to carry out research on its health effects, and that their failure to take concerted action makes them responsible for the fact that tobacco remains heavily promoted and a legal product in every nation on earth. Although it is true that tobacco production will ultimately have to be curtailed to eliminate this health menace, to attack those who brought the tobacco problem to light is bizarre. The need to deal with tobacco production in no way diminishes the importance of the many epidemiological studies that have illuminated the full detail of the problem. Skeptics might also question whether control of production is feasible without diminishment of consumption first; as the illicit drug trade illustrates, to believe that production can be turned off in the face of strong demand is simplistic.

All poverty is unacceptable We agree wholeheartedly that the study of the social causes of disease is an important epidemiological goal, and that societal causes can explain much of the variation in disease occurrence. We abhor tobacco promotion and production. We would like to see the eradication of poverty, and agree with Susser and Susser that epidemiologists should be well educated with regard to their public-health role. Nevertheless, the importance of societal causes of disease does not mean that biological pathways to disease should be ignored, or that epidemiologists who choose to 812

study causal mechanisms have been neglecting their mission. Furthermore, as with any public-health professionals who share humanist values, epidemiologists do not need to establish the health effects of poverty to know that society should aim to eliminate it. Two decades ago, a prominent college dean in public-health spoke out against nuclear war. His innovative argument was that nuclear war should be banned because it would disrupt the delivery of health services. In the same vein, today’s critics of epidemiology use health as an argument for social ends—such as the eradication of poverty—the desirability of which is obvious and quite independent of their public-health consequences.14 Perhaps the most valuable message in this new criticism of epidemiology is simply that those who wish to ease the burden of disease should not forget that the people of the world often bear larger burdens than those we sometimes choose to study. Nevertheless, epidemiologists cannot be expected to solve every problem, especially not those beyond our expertise. However well motivated, epidemiologists cannot rid the world of poverty. Even if we claimed that poverty is the root cause of all disease, which it surely is not, we would hardly be closer to solving the problem—just as we were no closer to eliminating the threat of nuclear war after pointing out that Armageddon would interfere with physicians’ treatment of their patients. We understand that adding voices to the chorus in favour of social betterment and human rights will help attract attention to fundamental social problems. There is no reason to believe that epidemiologists have become more reluctant to combat poverty, injustice, and oppression. Yet, epidemiologists are not social engineers; they are public-health scientists who have a right to specialise as they see fit. They should be free to choose the subject of their inquiries, whether it be social causes or molecular causes of disease. They should not be discouraged from answering questions such as whether vitamin-A supplements in poverty-stricken children will reduce morbidity and save young lives. The answers to such questions will prevent deaths with an immediacy that social reforms, however desirable, cannot match. To get such answers, we must study “decontextualised” disease mechanisms in individuals. It is remarkable that epidemiologists are now chastised for their scientific accomplishments, which include such victories as the elaboration of the effects of tobacco smoking on many diseases, and the effect of folic acid on neural-tube defects. Countless other fragments of useful epidemiological knowledge, such as the benefits of breast milk over infant formula, have enabled many people to improve their health even if they could not avoid poverty and repression. If an astrophysicist can study the origin of the universe without apology, should an epidemiologist have to apologise for work that is so practical? For helpful comments, we thank Katarina Augustsson, Cristina Cann, Nancy Dreyer, Sander Greenland, Stephan Lanes, Charles Poole, David Savitz, Ezra Susser, Mervyn Susser, Aleaxander Walker, and Allen Wilcox. Supported by grants from the Swedish Cancer Society.

References 1 2 3

Susser M. Epidemiology today: “A thought-tormented world”. Int J Epidemiol 1989; 18: 481–88. Editorial. Putting public health back into epidemiology. Lancet 1997; 350: 229. Shy CM. The failure of academic epidemiology: witness for the prosecution. Am J Epidemiol 1997; 145: 479–84.

THE LANCET • Vol 352 • September 5, 1998

ESSAY 4 5 6

7

8

Pearce N. Traditional epidemiology, modern epidemiology, and public health. Am J Public Health 1996; 86: 678–83. Beaglehole R, Bonita R. Public health at the crossroads: achievements and prospects. New York: Cambridge University Press, 1997: 120. Susser M, Susser E. Choosing a future for epidemiology, II: from black box to Chinese boxes and eco-epidemiology. Am J Public Health 1996; 86: 674–77. Plotnick G, Coretti MC, Vogel RA. Effect of antioxidant vitamins on the transient impairment of endothelium-dependent brachial artery vasoactivity following a single high-fat meal. JAMA 1997; 278: 1682–86. MacKay A, Rothman KJ. The incidence and severity of burn injuries following project burn prevention. Am J Public Health 1982; 72: 248–52.

9 10 11

12

13 14

Mouriño-Pérez RR. Oceanography and the seventh cholera pandemic. Epidemiology 1998; 9: 355–57. Colwell RR. Global climate and infectious disease: the cholera paradigm. Science 1996; 274: 2025–31. Sommer A, Tarwotjo I, Djunaedi E, et al. Impact of vitamin-A supplementation on childhood mortality: a randomised controlled community trial. Lancet 1986; i: 1169–73. Daulaire NMP, Starbuck ES, Houston RM, et al. Childhood mortality after a high-dose of vitamin A in a high risk population. BMJ 1992; 304: 207–10. Herrera HG, Nested P, El Amin A, et al. Vitamin-A supplementation and child survival. Lancet 1992; 340: 267–71. Wilcox AJ. From the Shy trial defense: a leak, or a red herring? Epidemiology 1997; 8: 684–85.

Essay

One fortunate old man

Michael Harmer In July, 1986, The Lancet published a personal paper entitled “Crohn’s disease: two fortunate young men”,1 which I had written with A G S (Joe) Bailey to celebrate our 50th and 43rd anniversaries since elective surgery for the treatment of so-called Crohn’s disease by excision of our affected small intestines. My operation was performed in 1936, when I was about to qualify from St Bartholomew’s Hospital, and Joe’s was done in 1943, when I assisted Mr Harold Wilson (as his then Chief Assistant) and who performed both our operations. We related these events in our article; and we promised ourselves that 10 years later we would write a follow-up paper. Ian Munro, The Lancet’s editor at the time, was generous enough to respond that he would look forward to us so doing. Fate decided otherwise. By 1996, I had been forced to submit to a couple of trivial complaints—an aortic valve replacement (from a small piglet) and a rather troublesome colostomy, which today behaves, possibly due to the absence of my ilea-caecal valve, more like an ileostomy. In 1997, Joe, alas, died. Hence my title. My purpose in writing yet again about this strange disorder, now universally known as Crohn’s disease, is to attempt to clear up, if possible, some misconceptions about its origin and nomenclature. First, the name. The landmark (sic) article published in the Journal of The American Medical Association in 1932,2 was written by three pathologists, since the surgeon who was also involved in the investigation, A A Berg, declined to put his name to a joint report. Crohn himself accepted that Ginzberg had done most of the work and never sought the eponym for himself. The practice of according authorship alphabetically led to the disorder’s present name. This I know because I spoke to and corresponded with Crohn and Ginzberg for several years before their deaths. Berg I never met. Lancet 1998; 352: 813–14 Perrot Wood, Graffham, Petworth, Sussex GU28 0NZ (M Harmer)

THE LANCET • Vol 352 • September 5, 1998

In actual fact, Crohn called the disorder “eelietis” (because of his New York accent) and that I found appropriate because Sir Kennedy Dalziel, the Glaswegian surgeon who had described the disease, in immaculate clinical detail, 20 years before Crohn and colleagues,3 had suggested that “the appearance of the terminal ileum resembles that of an eel in rigor mortis.” It was because I believed that Dalziel had not been given the credit due to him that I attempted to do so in 1987. My plea was turned down flat by the editors of both The Lancet and the BMJ, who argued that “Crohn’s disease” had come to stay and that was that. The Bristol editor took a more lenient view and published my suggestion4 that the name should be changed. Unfortunately Dalziel is pronounced Dayell, or something akin to that, and nobody could be expected to postulate such an eponym. So I suggested, as a compromise, Crohn-Dalziel disease; and to my surprise and gratification, I found it being used at a recent surgical conference. I once asked Ginzburg if I should attempt to change the name, “Not only you should (he replied); you must”. That is my justification. Second, the pathology. Crohn described regional ileitis as a clinical entity, and so it is.2 Contemporary pathologists have turned “the disease” into an academic beanfeast. Squinting down their microscopes they identify Crohn’s disease throughout the entire digestive tract, from the gingiva to the muscle of Ellis, better known as the Corrugator cutis ani. This is about as far from what Dalziel had in mind as calling Scotch, CocaCola. Next, history. Any searcher-after-truth knows that the further one looks back, the more references come to light. So although Dalziel anticipated Crohn and colleagues, he himself was certainly not the first. Moynihan, in 1907, and Mayo Robson, in 1908, had written papers on “Inflammatory tumours of the bowel”. I myself once suggested that, mouldering in the British Museum or the Bodleian, someone will find a 813