Bringing the Epidemiological Perspective to Pain Research: A Conversation with Gary Macfarlane

Editor’s Note: The first-ever North American Pain School (NAPS) took place from June 26-30, 2016 in Montebello, Quebec, Canada. An educational initiative of the International Association for the Study of Pain (IASP), Analgesic, Anesthetic, and Addiction Clinical Trial Translations, Innovations, Opportunities, and Networks (ACTTION), and the Quebec Pain Research Network (QPRN), NAPS brought together leading experts in pain research and management to provide 30 trainees with scientific education, professional development and networking experiences. Six of the trainees were also selected to serve as PRF-NAPS Correspondents, who provided first-hand reporting from the event, including summaries of scientific sessions and interviews with NAPS’s six visiting faculty members, along with coverage on social media. This is the first installment of interviews from the correspondents, whose work will be featured on PRF and on RELIEF, PRF’s new sister site for the general public.

Gary Macfarlane, PhD, is a professor and Clinical Chair in Epidemiology at the University of Aberdeen, where he is also Deputy Director of the Institute of Applied Health Sciences. He trained in statistics, computing science and medicine at the University of Glasgow before undertaking his PhD at the University of Bristol. He worked at the Division of Epidemiology and Biostatistics at the European Institute of Oncology in Milan before switching disease areas and leading a program of chronic pain research at the Arthritis Research UK Epidemiology Unit at the University of Manchester. His major research interests are in the epidemiology of rheumatic and musculoskeletal disorders, with a particular focus on pain (including fibromyalgia), fatigue and axial spondyloarthropathy (or ankylosing spondylitis). He also still undertakes work in cancer epidemiology. Macfarlane sat down in person with PRF-NAPS correspondent Pauline Voon, a PhD candidate in population and public health at the University of British Columbia, Vancouver, Canada, to discuss his path to epidemiological research, what epidemiology has to offer investigators studying pain, and his advice to those early in their careers. Below is an edited transcript of their conversation.

How did you become interested in the field of epidemiology, and in pain research?

I came through a rather unusual background. I did an undergraduate degree in mathematics and computing, and then specialized in statistics, at the University of Glasgow. During this time, one of the mathematics tutorial classes that I taught was taken by an epidemiologist, and he invited me to come work in the cancer epidemiology unit over the summer—and then suddenly I became interested in epidemiology. After that, I did a medical degree at Glasgow. Even once I’d done my medical degree, I could have decided to specialize in various clinical specialties, but I decided that epidemiology was for me.

The reason I decided that was because, when I was studying medicine, as well as being interested in treating patients, I was often asking the question: “Why has this patient got this particular disease at this time?” It was very much from the epidemiological perspective that I was thinking about this. After my PhD at the University of Bristol, I went to work in cancer epidemiology at the European Institute of Oncology in Milan, Italy. Then I came back in 1995 to the University of Manchester in the UK and worked for the Arthritis Research UK Epidemiology Unit. I was in charge of the pain program, having never undertaken any pain research at all, but that was my introduction to working in pain research.

What particular areas of epidemiological research interest you?

Epidemiologists, of course, are interested in etiology of disease. Within our Rheumatic and Musculoskeletal Disease Program, we’re doing epidemiological studies to identify new risk factors for chronic pain or to evaluate the strength of association between an established risk factor and chronic pain. I firmly believe that epidemiology has to inform intervention studies. When doing epidemiology studies, I’m thinking about modifiable factors, which could be the patient’s lifestyle, psychosocial factors or work environment, for example. The really interesting thing is that by focusing on modifiable factors you can then think about what the intervention would be to change them.

In terms of specific areas, a focus on fibromyalgia has for a long time been a major interest. It’s relatively common, and although we do understand something about its etiology and treatment, what we know is definitely inadequate. I have just chaired the European League Against Rheumatism working party to revise the guidelines for fibromyalgia management. Although we have lots of trials, all the treatments available have very modest effects, which has emphasized to me how much more we need to do to understand how we can improve outcomes for patients.

What are some of the challenges and advantages of studying pain using epidemiological methods compared to other approaches?

I mentioned that originally I was a cancer epidemiologist. One of the things that was evident to me when I moved to pain epidemiology is that the first major challenge is the definition of what you’re studying. We heard during NAPS, for example, that probably more than 80% of people in a population will experience back pain during their lifetime. I would suggest that this is not the phenotype we’re interested in, because that almost becomes part of normality. Many scientists think that emphasizing these very large prevalence figures provide a compelling rationale for funders and policymakers; I would argue the opposite. It does us a disservice to say, for example, that 80% or 90% of people will suffer from pain, because what we should be focusing on is disabling pain—pain that is severe, frequent and has an impact on people’s function and their role in society. This is probably around 20% of the population. So a clear definition of what we are studying, and its public health importance, is vital.

A second challenge is the natural history of pain, in the sense that it can be episodic; people have an episode of pain, which then resolves, and then they may have another episode. Epidemiologists traditionally rely on cohort studies, which recruit people who are free of the disease of interest, measure risk factors, and then follow them up. But it’s really difficult to find adults who are free of any experience of pain, and you could argue that they’re probably quite an unusual group anyway. Even in cohort studies, people who do not have pain currently will have experienced pain in the past, and that might have altered what they do, their lifestyle, and so forth.

An advantage, however, of pain being common is that when we ask patients to complete questionnaires or to come in for assessment, they can identify with pain; they think it’s relevant to them, and we know that’s something that influences whether people take part in epidemiological studies. Even if people have not experienced pain themselves, they probably know someone who has, and are therefore more motivated to take part.

Why is there less representation of epidemiological pain research compared to other areas of pain research, such as basic science?

That’s an interesting question, and I’m not sure that I’ve got the answer to it. When people go to medical school, they go really “to save patients’ lives”; they don’t go there with a population perspective. In terms of people’s exposure in medical school, they get drawn into particular clinical disciplines or particular scientific approaches, and perhaps we haven’t traditionally had the appeal to students that we need to have. I have found, however, that clinicians, having established themselves in a career, often find that they realize the importance of that broader perspective, and they come to the discipline later on to help them get it.

The one thing I’d say, particularly about the United States, is that whereas in other countries there can be suitable sampling frames to do epidemiological studies, and the health care system may be set up in a coordinated way to facilitate the studies, that isn’t really the case in the United States. There, it’s difficult to draw population sampling frames, and the organization of health care is such that patients can go to different providers. Whereas, for example, in public systems like the United Kingdom, there’s a single gatekeeper, and a common record system, and that can make it easier to do the studies. That also contributes to why there are probably more epidemiological groups studying pain and rheumatic disease in Europe (particularly in the United Kingdom) compared to North America.

What are some of the topics or methods in epidemiological pain research that you’re most excited about?

At NAPS I talked about the case-crossover study, which although not new to epidemiology is relatively new to pain research. It involves using people as their own controls, which really means we can have a tighter control of confounding. It’s been used in studies of back pain, and it’s been able to demonstrate associations with physical factors, including physical load and injury; these associations have been much stronger than we’ve been found before.

The other study type that is really exciting is called Mendelian randomization. This involves looking at the genotypes of individuals that influence their level of exposure to a specific factor. For example, we know there are some genetic influences on body mass index. What Mendelian randomization does is ask, for example, if I identify people with a genotype that is linked with a higher body mass index, is that genotype related to back pain? If it is, that provides very strong evidence not only of association, but perhaps of causality. We’ve done some studies suggesting that body mass index and chronic pain may actually be causally related rather than just confounded by other factors.

In your research, or in the work of other investigators, have you found any particular factors that have been surprising to you in terms of their confounding role in pain?

That’s a really tough one. One of the things I’ll point out is something I was surprised at, and we still don’t understand the relationship: we have shown that people with chronic pain have premature mortality. We’ve been trying to understand why that is, in particular by looking at lifestyle factors. We started off by looking at levels of physical activity, because it’s entirely possible that people who have developed chronic pain become less physically active, which puts them at risk for a whole variety of diseases, including cardiovascular disease and cancer.

It appears that physical activity does play a role, but something that may be even more important is diet. It’s been relatively under-researched, but it does appear that people with chronic pain have markedly poorer diets than people who don’t have chronic pain, and this poor diet may be a factor in terms of an increase in their risk for cancer. I can’t be certain about that, however—perhaps it’s confounded—but that’s what we’re looking at, at the moment. Alternatively, it may be that diet is on the causal pathway—patients develop chronic pain and because of that they feel less inclined to spend time making and preparing healthy food options. Although it’s not startling news that diet and physical activity are related to mortality, my understanding is that in terms of managing pain patients, the role of lifestyle factors, such as diet, is mentioned in very few consultations. Having patients you’re managing for chronic pain with you in the consulting room may be a really good opportunity to say to them that you understand chronic pain can sometimes affect their diet, willingness to cook, and willingness to eat healthily, and it could be an opportunity for us to intervene, to prevent the increased mortality that seems to occur. But that’s still an open question.

A question people often ask me about my own research is whether I am studying physical pain or emotional pain. It’s often very hard to separate those things, especially in observational or self-reported research in epidemiology. Do you have any strategies to address this?

This is an issue for us as well. Often when I present my research and show the importance of psychosocial factors, patients think that what I’m saying is that their pain is “all in their head” and not real. Certainly, epidemiologically, we’ve moved towards thinking that both components will be there and are very important, in most people. It’s important to emphasize that even if, for example, there’s an obvious physical injury causing pain, the brain has a very important role in how much you feel, so even people who have the same physical injury will feel pain quite differently. Part of that will be due to psychological factors, so it’s a false dichotomy between physical and psychological pain.

Looking back at your career trajectory, if you were to give a piece of advice to trainees about something you found to be most helpful early in your career, what would it be?

Take advantage of opportunities as they arise. Not every opportunity comes for you at the right time, but always position yourself so that you are able to take advantage of opportunities that come along.

The other major piece of advice is to try to seize the opportunity of working in a country different from where you have trained. Things get done in different ways in different places; you work with different people, and with different approaches to research. All that will help you fashion your own experiences, so when you become a principal investigator, you can draw on them and choose the best parts of your various experiences—and discard the worst!

Can you describe your applied epidemiology course?

We have run now, for 24 years, an annual course called the “Intensive Course in Applied Epidemiology,” which very much has a focus on pain and rheumatic disease. What we try to do is cover a master’s level module in epidemiology in one week. We’re giving people skills in terms of designing studies, analyzing data and interpreting the results. This can be useful both for doctoral students who are starting out on their career, and also for people with more established careers who then feel that these techniques may be really important to them. People come from all over the world—really bright people. We have a great week, as a residential course, learning methods that hopefully will be of use to them, not just in epidemiology, but in conducting research studies generally.

Thanks for speaking to PRF.

Thank you.

Additional Reading:

Alcohol consumption in relation to risk and severity of chronic widespread pain: results from a UK population-based study.
Macfarlane GJ, Beasley M
Arthritis Care Res (Hoboken). 2015;67(9):1297-303. [PMID: 26212017]

The prevalence of fibromyalgia in the general population: a comparison of the American College of Rheumatology 1990, 2010, and modified 2010 classification criteria.
Jones GT, Atzeni F, Beasley M, Flüß E, Sarzi-Puttini P, Macfarlane GJ
Arthritis Rheumatol. 2015;67(2):568-75.

The re-evaluation of the measurement of pain in population-based epidemiological studies: The SHAMA study.
Flüß E, Bond CM, Jones GT, Macfarlane GJ
Br J Pain. 2015;9(3):134-41. [PMID: 26516569]

Alternative population sampling frames produced important differences in estimates of association: a case-control study of vasculitis.
Macfarlane GJ, Jones GT, Swafe L, Reid DM, Basu N
J Clin Epidemiol. 2013;66(6):675-80. [PMID: 23313107]

Diet, lifestyle and chronic widespread pain: results from the 1958 British Birth Cohort Study.
Vandenkerkhof EG, Macdonald HM, Jones GT, Power C, Macfarlane GJ
Pain Res Manag. 2011;16(2):87-92. [PMID: 21499583]

Musculoskeletal pain is associated with a long-term increased risk of cancer and cardiovascular-related mortality.
McBeth J, Symmons DP, Silman AJ, Allison T, Webb R, Brammah T, Macfarlane GJ Rheumatology (Oxford). 2009;48(1):74-7. [PMID: 19056799]

The epidemiology of chronic syndromes that are frequently unexplained: do they have common associated factors?
Aggarwal VR, McBeth J, Zakrzewska JM, Lunt M, Macfarlane GJ.
International Journal of Epidemiology. 2006 Apr 1;35(2):468-76. [PMID: 16303810]

Other Resources:

Intensive Couse in Applied Epidemiology
http://www.abdn.ac.uk/iahs/research/epidemiology/icae-aberdeen-course-158.php

European League Against Rheumatism – Recommendations for management
http://www.eular.org/recommendations_management.cfm