Most epidemiologists will know that the occurrence of many diseases is driven by complex social factors and political conditions operating through health behaviours, exposure to risk factors, and ultimately through molecular mechanisms. In this stream of causation many epidemiologists are rafting quickly downstream towards molecular epidemiology or nano-epidemiology (at least the internet domain name has been taken). Research funding is downstream from seeking the primordial social and political causes of diseases which requires the study of macrostructures but is focusing on understanding the micro cosmos of biology. The reasons for these shifts may be good or bad – not only the devil, but also the money, is in the detail. Using Aristotelian terminology we are neglecting studies on the final and efficient causes to study the material and formal causes and that will come at a price.
First, it is not certain that our epidemiological toolbox is useful for this new line of research. Most of what we base our current understanding on today in “modern epidemiology” is the risk factor epidemiology from the middle of the 20th century. Complex biological feedback systems, pathogenesis based upon signalling between different genes, and more may need new tools for pattern recognition.
Second, and more importantly, we may throw the baby out with the bathwater. What we have achieved was our ability to identify preventable causes of diseases and contribute to the evaluation of public health interventions for their control. Not many epidemiologists seem to be concerned about the speed of this transition, or to be concerned about our ability to continue to deliver equally important findings for the public health as epidemiologists did in the “risk factor” epidemiology period. Funding in the US and Europe goes with the newest technology and old virtues in epidemiology die out as fast as ice is melting away in Greenland. Notice, if you read funding applications, how few epidemiologists even address the issue of whether their hypothesis makes epidemiologic sense in terms of its coherence with observational epidemiology. Does disease occurrence over time or between populations follow a path that is compatible with the putative exposure distribution? Notice how seldom you see this addressed in epidemiologic papers, even in those published in epidemiologic journals. Notice how little even epidemiologists nowadays take an interest in whether their findings can be translated into public health practice. Read Morris’ “Uses of epidemiology” and then turn to contemporary textbooks in epidemiology and notice how the history, the philosophy and the public health aspects have disappeared. Or check the changes in the definitions of epidemiology to see how public health and populations have disappeared from the definitions. What remains is only “to study determinants of diseases”.
We have come a long way from Virchow’s statement that “politics is nothing but medicine on a big scale” to Pearl’s “causes have now been mathematized”. Is that good or bad? Or are we prepared to take this step? Should we welcome it or should we value that part of epidemiology that has its roots in modifiable social conditions and environmental exposures? Should we insist that we have the right (perhaps even the obligation) to study associations between external exposures without being guided by biological theory; to insist that epidemiology is basic science and not just a translational science from biology to public health practice?
The transition to nano-epidemiology could be seen not only as an attempt to follow contemporary hype but as an adjustment to the fact that all important risk factors have been identified and what now remains is to discover the mechanisms of pathogenesis. Unfortunately, this is not true. We are only just beginning to collect data over long time periods that allow us to study diseases in a life course perspective.
Furthermore, most epidemiologic studies have addressed cancer and cardiovascular diseases. Relatively little epidemiologic research has been devoted to many other diseases, including infectious diseases. Although many infectious diseases have a necessary cause (by definition), the other component causes determine whether the exposed fall ill or not.
Epidemiologic research and methodology has, of course, to evolve in order to meet new challenges, but it is important not to forget that one of the old virtues, describing and understanding disease patterns in populations, should not be forgotten. In spite of microscopes and new laboratory equipment the microbiologists could not explain why the infectious diseases took the course they did in the population. Epidemiologic research survived a take-over by microbiologists in the beginning of the 20th century because microbiologists had no way of understanding the big drivers of the epidemics in populations. Epidemiologic research may also survive molecular biology, not by becoming nano-epidemiology but by keeping in touch with how diseases, illness and health develop over time, between populations and within populations; diseases are not just results of accidental or random mutations or lesions.
The IEA will be an advocate for epidemiologic research linked to preventive medicine and public health. Being a society with representatives on all continents we have an obligation to take an interest in the most important determinants of health. It is encouraging and promising that private foundations like the Bill and Melinda Gates Foundation also give priority to this area of public health. It calls for training more epidemiologists that are willing to work in underprivileged areas and to work with people who have a long way to go before they catch up with health conditions in affluent societies.
This research should of course use modern biologic techniques to advance our knowledge when needed. Monitoring emerging diseases may well involve mass screening of the infectious agents’ genetic drift linked to health responses in animals and humans. Understanding the epidemics of obesity, diabetes and many other diseases does involve studies of genetic factors and epigenetic modifications in fetal life, but with a focus on the factors that are modifiable. The control of the SARS epidemic globally is a good example of using both contract tracing and gene sequencing to stop the epidemic within a short time which would not have been possible a century earlier. Likewise, in Avian flu, knowing the variation in H5N1 viral strain, could be used to trace the mode of spread (i.e., spread from central Thailand or from China). However, the definitive means of control of Avian flu remains through quarantine, and changing the behavior of people who raise poultry and the ways in which the food industry handles animals.
The success story of how epidemiologic research made it possible to understand the causes of cervical cancer and ultimately control the disease illustrates well the leading role epidemiology has in preventive medicine. By identifying risk factors and geographical distributions it became clear that the disease had an infectious basis that spread through sexual contacts. Identification of the virus and then a possible vaccine provided additional and important means of preventing the disease. The story illustrates that epidemiologic findings precede biological understanding. The same applies to the early epidemiological observations on Kaposi’s sarcoma in the 1980s among homosexual men of the syndrome that we now know as HIV/AIDS.
The cervical cancer story also illustrates that it is important to continue doing explorative frontline epidemiologic research that need not be linked to biological theory. Epidemiology is not only a scientific discipline devoted to corroborating biological theory but a science that generates hypotheses for the biological field. Just think of the studies on smoking and lung cancer in the 1950s that at that time was not initiated by a biologic theory of carcinogenesis, or the studies on EMF or non-ionising radiation. It may well be that epidemiologic research will indicate causal paths that at present we know nothing about. Although advances in our biological understanding have been impressive this understanding is still in its infancy and exciting new discoveries will in the future not only add new knowledge but also show that many of the current theories and facts have a short life expectancy.
Epidemiologic research will take part in this process especially if we pay attention to how diseases are distributed in the population over time, between places and persons. Better monitoring data and expanded surveillance using methods that have been applied in common cancers and cardiovascular diseases to other diseases are needed. We should use data on who are diseased, how the frequency of the disease evolves over time and place to examine hypotheses on possible causes. We should not accept that we always need to have a biological explanation at hand – indeed biological plausibility is one of the weaker “causal” criteria, none of which are “sine qua non” criteria except, perhaps, temporality.
Epidemiologists, like other scientists, have to follow what funding agencies are interested in funding but these funding agencies get their inspiration from scientists that lobby for their field of research. Epidemiologists need to be better actors in this game through our organisations and as individuals when we have the opportunity. Let’s end this New Year greeting with a hope also for the need to be organised to get influence.
All the best wishes for 2007,
Jorn Olsen, Shah Ebrahim, Chitr Sitthi-amorn