Cognitive Dissonance and Evidence Based Medicine - David Naylor’s Kergin Lecture

Frederick Gordon Kergin Frederick Gordon Kergin was born at Port Simpson in British Columbia in 1907. He began his studies at the University of Toronto at age sixteen and graduated from the Biology and Medical Sciences program in 1927. In 1931, Kergin became a Rhodes Scholar and spent the next two years at Oxford University in a Master’s Degree program in physiology and anatomy, graduating with first-class honours. In 1934, he began the four-year Gallie Course in surgery at TGH and obtained the fellowship of the Royal College of Surgeons of England in 1935 and that of the Royal College of Physicians and Surgeons of Canada in 1939. In 1937, he joined the surgical staff of Toronto General Hospital and later took the role as Chair of the Department of Surgery of the University of Toronto and Surgeon-in-Chief of the Toronto General Hospital from 1957 to 1966. He was a pioneer of thoracic surgery in Canada and served as President of the American Association for Thoracic Surgery. In1966, he was appointed Associate Dean in the Faculty of Medicine and was responsible for developing a new undergraduate curriculum and planning the conversion of Sunnybrook Hospital to a teaching institution with full-time faculty. Dr. Kergin chaired the editorial board of the Canadian Journal of Surgery for many years and served as a trustee of the R.S. McLaughlin Foundation. His major contribution to the University was in education, particularly in structuring the residency programs such that an integrated program amongst all fully affiliated hospitals was established. Professor Kergin died in 1974, a man with eclectic interests that included teaching, research and university administration, as well as many outside of medicine. (http://surgery.utoronto.ca/events/kergin-lecturers.htm)

James Rutka and David Naylor

The 2017 Kergin Lecture was delivered by David Naylor, former Dean of the Faculty of Medicine and President Emeritus of the University of Toronto. Naylor’s theme was ‘Cognitive Dissonance and Evidence-Based Medicine’ [EBM]. Based on his studies of an American religious cult, psychologist Leon Feininger coined the term ‘cognitive dissonance’ in 1956 to designate the distress people feel when reality conflicts with deeply-held beliefs. Cognitive dissonance is resolved by denying, ignoring, or reinterpreting the contradictory finding which does not fit with our biases or beliefs. Naylor drew on his long experience as a researcher and policy advisor to examine applied health research, healthcare policy-making, and clinical reasoning, all viewed through the lens of cognitive dissonance.

Starting with research, Naylor observed that clinical epidemiology emerged in the 1970s and 1980s as a discipline that aimed to enhance the rigor of clinical studies and bring research results to bear more fully on clinical decisions. These insights were synthesized and presented to the profession in the early 1990s as ‘Evidence-Based Medicine’ [EBM]. EBM was described as a new clinical ‘paradigm’, i.e. a system of assumptions, concepts, values and practices that constitutes a way of viewing reality. It characterized clinical experience as an unreliable source of evidence – an apparent devaluation of clinical judgement that understandably unsettled many surgeons given the highly case-based nature of their work. Naylor observed that, more generally, EBM as a paradigm has continued to struggle with the dilemma of the applicability of evidence and the unreality of the ‘average patient’. Small effects that turn out to be statistically significant in giant randomized trials mean that many patients are exposed to side-effects for everyone who benefits from a given ‘evidence-based’ treatment. EBM acolytes often argued that the solution was to stratify subgroups of patients by baseline risk, assume the same relative benefit would accrue to all, and therefore infer that the highest-risk patients would gain the most in absolute terms. However, studies using sophisticated biomarkers are starting to overturn this mode of reasoning.

To illustrate this point, Naylor showed us a randomized trial of the cardiovascular drug Dalcetrapib (Tardif JC et al, Circ Cardiovasc Genet. 2015;8(2):372-82) that initially found no difference when the drug was compared to placebo. Use of biomarkers later revealed that the drug was highly beneficial in one genetic subgroup and harmful in another. The two effects cancelled, leading to an erroneous ‘evidence-based’ conclusion. Naylor suggested that this result was a bellwether for the challenge facing the current incarnation of EBM as ‘the medicine of averages’. He observed that a counter-paradigm was emerging as biomolecular characterization of patients continued to advance, thereby enabling better tailoring of treatments. He predicted that new molecular markers and other measures such as functional imaging were likely to compel reconsideration not just of who received specific treatments, but how we define disease entities, particularly in disciplines such as psychiatry, with its descriptive Diagnostic and Statistical Manual of Mental Disorders (DSM).

Naylor emphasized that consideration of variations in patient characteristics and anticipated responses to treatment has been an integral part of expert judgement dating back centuries in clinical medicine. What is different now is the convergence of our deepening understanding of human biology with other factors such as the use of digital devices to enable continuous monitoring of patients with sophisticated sensors, improved imaging, automated treatments based on digital monitoring and the application of artificial intelligence. Sophisticated tissue engineering techniques may transform not just the field of transplantation but surgery in general. This layering of diverse disruptive forces has meant that the once-popular term, ‘molecular medicine’, is already being largely supplanted by terms such as ‘personalized’ or ‘precision medicine’.

Just as the emergence of EBM seemed to engender cognitive dissonance among those attached to other modes of thought and action, so also was it now ironically the case that EBM fundamentalists were among the most vocal critics of personalized or precision medicine. Naylor cautioned, however, that personalized or precision medicine was far from a panacea. It had the potential to provide remarkable improvements over the “shot-gun approach” of EBM, but many exaggerated claims were already being made for this latest paradigm and the potential costs and risks are enormous.

Other countries such as the UK and Australia were more enthusiastic about precision medicine, and more thoughtful about developing a reliable knowledge base and strong framework for funding and using these concepts in practice. Developing a Canadian national strategy for personalized medicine was accordingly among the recommendations made in 2015 by a distinguished Advisory Panel on Healthcare Innovation that Naylor chaired for the federal government (link to PDF). Although the Conservative Government of the day shelved the report, it has found new life under the current Liberal Government, underscoring Naylor’s comment that political ideology of all types carries its own forms of cognitive dissonance. Naylor among others who promoted EBM in the 1990s had emphasized that clinical decisions would continue to rest not only on evidence, but on a given patient’s values or preferences and the context of the clinical encounter (Naylor CD. Lancet 1995; 345 (8953):840–2). He wondered if that list should now be expanded to include cognitive psychological factors, and particularly in the realm of healthcare reform, overtly political or ideological considerations.

In support of that point, Naylor reviewed some work from his early years at the Institute for Clinical Evaluative Sciences (ICES), showing remarkable variation in rates of caesarian sections, hysterectomy, knee replacement, and breast conserving operations. This work had caused a media sensation when it first appeared in 1994. Politicians were quick to criticize the profession; physicians and surgeons in turn rushed to explain away the variations in very creative ways.

Naylor emphasized that while some practice variations reflected indefensible departures from rigorously assessed practice standards, in other instances they reflected evidentiary uncertainties, different financial and organizational contexts, and regional or national clinical cultures. On the latter point, he reminded the audience that expert panels from different countries would arrive at different views about the appropriateness of surgery when given the same evidence and patient case scenarios. Naylor then showed real-world examples of this phenomenon in the realm of differences between Canadian and American practice patterns in use of cardiovascular procedures after myocardial infarction (Mark DB et al. N Engl J Med 1994; 331:130-135).

Naylor also summarized several studies led by Toronto researchers illustrating how errors in cognitive processing affect decision-making. For example, physicians were more likely to favour testing and treatment when considering their recommendation to an individual patient than when they were asked to consider how they would write guidelines for a group of similar patients (Redelmeier DA, Tversky A. N Engl J Med. 1990;322:1162-4). The framing of treatment data also has a powerful effect. A new drug might reduce death rates for a given condition from 4 per 100 patients treated to 3 per 100 -- a 25% relative risk reduction. However, if the same data are shown as a 1% absolute reduction, or represented fairly by the statement that 100 patients must be treated with the drug to save one life, then physicians become much more cautious about recommending the new medicine (Naylor CD, Chen E, Strauss B. Ann Intern Med. 1992;117(11):916-21).

As a final example of how politics and cognitive dissonance can shape the use of evidence, Naylor pointed out that in the early 1980s a rigorous randomized trial undertaken by RAND researchers had shown that costs were lower and outcomes similar with a comprehensive capitated plan (then known as an HMO and now more commonly termed an integrated delivery system) as compared to Canadian-style health insurance (Manning WG et al. N Engl J Med 1984;310:1505-15). The findings had limited uptake in the US due to lobbying by organized medicine and the private insurance industry, and were downplayed here because of smugness about Canada’s superior healthcare system and physician unease about changes in compensation modalities. In part because of our refusal to embrace such changes, the performance of Canada’s healthcare system is now seen by many experts as lagging behind a number of OECD peer nations.

Naylor closed the Kergin Lecture with two aphorisms that encapsulated his theme of cognitive dissonance and evidence-based medicine/policy-making. The first was from a collection of essays on medical history published in 1991: “The enduring lesson of history may be that social change is inevitable and institutional progress possible, but human nature is wonderfully intransigent” (1). The second, arising from three decades of experience as reflected in the lecture, was shorter: “How we think is more important than what we know --- or think we know”.

(1) Naylor CD, ed. Canadian Health Care and the State. Montreal: McGill-Queen’s University Press, 1992, p12.