A conversation with Dr. Jack Siemiatycki, epidemiologist

Maureen Lafrenière

Dr. Jack Siemiatycki has a Masters degree in Statistics and a Ph. D. in Epidemiology, both from McGill University. A former National Health Research Scientist of Health Canada’s National Health Research and Development Program, and Distinguished Scientist of the Medical Research Council of Canada, Dr. Siemiatycki is currently a Canada Research Chair and the Guzzo Chair in Environment and Cancer at l’Université de Montréal.

He has served on over 100 national and international expert advisory bodies, on many grant review panels, and on editorial boards of the American Journal of Epidemiology and many other journals. Dr. Siemiatycki has authored or co-authored over 150 peer-reviewed articles, numerous scientific reports and presentations, and also co-authored children’s books.

I met with Dr. Jack Siemiatycki in July 2009 to ask about his work in cancer prevention.

We all like to hear how people find their careers: what led you to work in disease prevention?

I came out of the 60s generation and, leaving university, I stumbled into epidemiology by accident as a summer job, and I’ve stayed with it ever since. It struck a chord as something socially relevant, something that had the potential to do good in society: to study the determinants of disease, and therefore to be able to prevent disease. It was also something that I found easy, and it seemed to provide an adequate income, so that was like the cherry on top.

For a very long time, epidemiology has been a marginal specialty within biomedical research; marginal in research dollars, in the number of people that went into the area and how they got into it. Nineteenth-century physicians who took an interest in infectious [‘epidemic’] diseases eventually developed, after the Pasteurian revolution, a model to study them, which led toward the disciplines that we now know as microbiology, bacteriology, and immunology. Epidemiology was developed in a kind of weird, almost artisanal way in the early 20th century in England and the United States, mainly by individual physicians carrying out research and developing the methods.

Modern epidemiology can be said to have begun in the 1950s and 1960s, probably spurred by research concerning smoking and lung cancer or cardiovascular disease in relation to blood pressure and cholesterol. The field has remained marginal, a sub-specialty for MDs but also an area for non-physicians as it draws on many streams of expertise, including medicine, pathology, statistics, sociology, and – depending on the kind of research – psychology, environmental sciences and others.

I got into the field at an excellent time, as it was just starting; McGill’s program was the first and so I was the third Ph. D. in epidemiology at McGill and in Canada, and environmental issues seemed like the core business of epidemiology to me. It was socially relevant, interesting, a growth area, and a lot of people going into the field in those early years were similarly motivated by social concerns.

In the intervening years, there have been a lot of changes, and not all of them happy; the environmental focus – a mainstay from the 60s to the mid-80s – has largely disappeared. Most epidemiology now deals with things like genetics, post-market surveillance of drugs using epidemiological methods, clinical epidemiology (to address problems of diagnosis and treatment of disease) and, in Canada over the past ten years, there has been a particular development because of the advent of the Canadian Institutes for Health Research (CIHR). Also, much of what is considered public and population health has been co-opted by social science researchers, and this has further marginalized epidemiological and environmental research. So environmental research has been on a decline in Canada for the last 20 years.

Is there a correlation between available funding and that decline?

It’s a combination of things, a sort of vicious circle, with a feedback loop between funding and researchers: if there’s more money in an area, people go into it, and if there’s less money, people get out of it. With environmental epidemiology, there isn’t a lot of money or many people doing it. There are not many people partly because there’s not much money for it and there isn’t much money partly because there aren’t a lot of people doing it.

That’s quite remarkable, given the level of public awareness about environment.

It’s astounding; people don’t believe me. When I explain this, even those in the biomedical milieu don’t believe me. I have specifically gone out and analyzed the kind of research topics that are supported by the public funders in Canada like the CIHR and the National Cancer Institute of Canada (the research arm of the Canadian Cancer Society), or other cancer charities in Canada, and it’s incontrovertible. The proportion of the total cancer research budget devoted to environmental research is somewhere between 1% and 3% of the total cancer research budget in Canada.

It’s a great deal less than one would guess.

It’s much less than most people guess and it’s much less than what people expect when they give every year to charity or when they pay their taxes expecting the government is funding cancer research. I think they expect that environmental agents should be perhaps not the top priority, but one of the priorities.
I have a very broad concept of environmental cancer – I almost mean it to include everything except the genetic baggage we’re born with. It includes diet, and not just the pollution of our food by pesticides or other contaminants, but also the nature of our diets (fat, fibre, etc.). It certainly includes pollutants in the air, water and so on, but also includes our work exposures and work habits, whether there is sleep deprivation, whether there is second-hand smoke, psychological stress (at work or at home), being able to block the electromagnetic environment – I think all of these things are part of the environment – the psychological, the physical, the chemical. So putting together all of that, and not just what people conventionally think about as the environment (air, water and soil), my estimate of 1% to 3% is very low – it’s purely lip service.

But epidemiology plays a particular role in researching disease and its prevention.

As far as I’m concerned, epidemiology is the basic science of public health and without epidemiology, you can’t target anything. You can’t figure out if you should advise people to reduce their exposure to sun, or increase it; you can’t advise them to eat more of this or more of that; you can’t advise regulatory agencies to lower levels of a certain pollutant; you can’t advise industry whether they should phase out one type of chemical or another; you can’t advise people even to stop smoking if you don’t have the epidemiology underlying it.
All public health action flows from good epidemiology, but there isn’t much being done in Canada. We’re at the stage where public health in Canada is acting on research results that were produced by epidemiology in the last 50 years (e. g., effects of smoking, excess sunlight, asbestos, high blood pressure, and water fluoridation). There is virtually no new research going on in epidemiology to produce new basic knowledge on which future disease prevention can be based.

When you spoke of 20th-century epidemiology and of tobacco-related and cardiovascular diseases, those were newly emerging diseases, without wanting to use the word ‘epidemic’.

It was an epidemic. If you look at it on a scale of decades and graph lung cancer incidence, you will see that it was quite a rare disease until the beginning of the 20th century; then suddenly, it rises dramatically and peaks around the 1970s, then starts to decline. Apart from the time scale, that graph looks like a typical epidemic graph. So there has been an epidemic of lung cancer in the 20th century. It’s the time scale that differs from an epidemic of cholera or typhoid.

We saw some restrictions introduced last year [in Canada] on bisphenol-A; that must have had an epidemiological basis.

Actually, it’s an excellent question and in a way, it challenges some of the things I said about epidemiology being the basic science because it’s not the only basic science of public health. Toxicology is also an important science and this is an example. There are some topics that cannot be studied well using using epidemiology, and where the best that we can do is to study them in animal models.

Is that because it can’t be reduced to one variable, or is it something else?

It’s often because we can’t identify adequate human populations that have been exposed to a given agent and delineate the exposed from the unexposed, and the degree of exposure. Smoking was a low-hanging fruit; it was an easy one to get a handle on. Exposure assessment is actually a kind of handmaiden of epidemiology – it’s an area of science that will determine whether epidemiology is feasible or not.

If you can identify groups of people that have been exposed using reasonable markers (questionnaires, blood tests), then you have a fighting chance of using epidemiology to address the question. However, for cancer, there’s typically a long lag period between exposure and disease. Even if you had a marker – blood, urine, or a chemical metabolite – they’re often markers of an exposure within 24 or 48 hours, which are not useful in telling you what that person was exposed to 10 or 20 years ago.

In those circumstances, we often rely on animal experimentation as a proxy for human evidence. It’s not a perfect proxy; even between mice and rats, the concordance in biologic responses is far from perfect, so between rodents and humans, we know there must be a lot of discrepancy in response. But when that’s all you have, using the precautionary principle, if a chemical showed carcinogenicity in rodents, then maybe that’s where we should focus our concern and our regulations.

Another issue is the talk about intergenerational after-effects of exposures, either in utero, or exposures during the lifetime of parents that might eventually have an impact on, or create susceptibility in offspring.

It’s an extremely difficult area to study in humans. By definition, it would take generations to investigate. It has been studied to some extent in animal models and it’s been shown that certain chemicals do have effects on future generations. How that plays out in human terms is not well known.

You mentioned other obstacles in carrying out epidemiological research apart from funding.

One of the most annoying obstacles is the culture of privacy and confidentiality that has emerged over the last 20 years, culminating in regulations concerning access to data about human subjects, as well as ethics committees in hospitals and universities that have become more and more highly structured. There are more committees and entities that one has to deal with, they get increasingly bureaucratic with time, and their standards invariably rise.

In my view, the benefits are far fewer than the costs – not only monetary, but significant costs that society bears. There are social costs to research not being done, and sometimes not even being undertaken because researchers say, “This is hopeless; I don’t want to spend a year of my life trying to get this through a committee. ”

Within my own research team, we have probably submitted between 50 and 100 applications over the past 30 years to various kinds of ethics committees, so we’ve had a lot of experience over a long period of time. In the early- and mid-80s, we tended to deal with people on the basis of common sense, whereas we now deal with some who have strong, ideologically-driven and bureaucratically-driven imperatives to protect their conception of human dignity at all costs.

And the logical conclusion, when you work through that rationale, is that protecting someone’s privacy is more important than protecting the health and lives of their children and grandchildren. If given an informed choice: “Do you want to have legitimate medical researchers have access to information about you which can save your life, or can save your children, your grandchildren and future generations for the cost of letting them access confidential information about you in a way that guarantees that they maintain the confidentiality, or would you prefer to not allow anyone access to any information about you and therefore forego any future possible be-nefits of research? ” I can’t imagine that more than half of one percent of the population would opt for the latter.

But those in the ethics industry are precisely the ones who value privacy above everything else; they decide what kind of research gets done, and what doesn’t. If they are non-smokers, and you confront them with the fact that our understanding of [the dangers of] smoking came from the kind of research that they are now prohibiting, how do they square the circle that they are benefiting from the fruits of research done on people 50 years ago – with unrestricted access to information – but are not willing to share information that would allow other people to benefit?

When it comes to disease prevention, concerns have been raised about public education on cancer prevention focused mainly on lifestyle risk factors such as diet, physical activity or tobacco, largely ignoring risks associated with environmental exposures that most of us have little or no control over. Should research on the latter be a priority for public funding?

I think there is a perception in much of the scientific community (biomedical community) that the important risk factors for cancer are individually controlled, or lifestyle ones. I think that is partially founded and partially unfounded. Namely, that among the ones that we know about and understand, (i. e., the cancer risk factors that have been well studied using epidemiological methods), the ones that tend to produce the highest ‘population-attributable risk’ are those that combine potency with a prevalence of exposure in the population.

Smoking, for example, is both a very prevalent characteristic in the population, and a very potent carcinogen. Those combine to create many cases of disease, and also produce diseases that are quite prevalent, like lung cancer, bladder cancer, and others. If you look at vinyl chloride monomer, which is a rare exposure and which causes a fairly rare disease, even though it’s very potent at causing that disease, the fact that it’s a rare exposure and produces a rare type of cancer significantly limits the total number of people who will be afflicted as a result of that exposure.

So if you compare smoking with vinyl chloride, smoking is way more “important” as a social cause of cancer. There are a number of other environmental-type carcinogens which have been studied and characterized, but when you look at physical activity, sunlight, or smoking (the ‘lifestyle-determined’ things), the attributable risk tends to be higher for those because they tend to be more prevalent exposures than some of the environmental ones that have been identified. The problem with that logic is that [lifestyle risk factors] have been identified because they’re much easier to identify. Identifying carcinogens or other pathogens in the general environment (in air and water pollution, or in food) is so difficult to do that at best, we have a very tenuous handle on whether there is a causal association between these exposures and cancer or other diseases, and if so, what the magnitude of that association might be.

We have much less certainty about environmental risks than we have about some of the more easily studied factors, like alcohol, smoking, and physical activity. Too many people draw the conclusion that if something hasn’t been identified as important, it’s therefore not important. And without the funding, the expertise and the interest to study them, we may never discover hard evidence for other important environmental carcinogens, and therefore, it will always be an elephant-to-a-mouse comparison, when you compare what we know based on what we have studied.

It then comes down to knowns vs. unknowns: if I know that consuming a certain product has a risk, then I will modify my behaviour, but if I never hear about the risk, I might just continue as before.

To comment on the way you phrased the premise (lifestyle vs. environmental exposures), I would never say that it’s a question of there being too much research on lifestyle factors and not enough on environmental factors, because there isn’t enough known about lifestyle factors. The whole ‘pie’ of epidemiological research is tiny, and to divide ‘tiny’ into sub-sets, I wouldn’t say that one has too much; I would say that it all gets too little. So I would go on looking at the other 97% of the cancer research budget and say about all the research that has been carried out over the last 50 years: what has it actually given us for preventing cancer? The war on cancer has been fought on that [other] kind of research, as opposed to whether this or that pesticide is dangerous, or whether a certain lifestyle is a cancer risk.