Evidence-Based Psychiatry: The Pros and Cons
Maurice Dongier, MD, FRCPC, Guest Editor
The development of medicine as a science only dates back to the 19th century. But Lesage, Stip and Grunberg (1) remind us that Galen, nearly 2000 years ago, had already identified 2 cultures among physicians—the “empiricists” (evidence-based) and the “rationalists” (also called “subjectivists”). Clearly, the methods of the empiricists are the very distant ancestors of modern evidence-based medicine (EBM), even though the current criteria for “evidence” are more defined.
An editorial in the British Medical Journal (2) has summarized “what it is and what it isn’t,” and we are privileged to benefit from a contribution to the present issue by John Geddes and Stuart Carney (3). John Geddes is the director of the Centre for Evidence Based Mental Health and editor of its electronic journal (4).
The Lesage, Stip, and Grunberg paper presented in this issue is based on a symposium organized at a recent annual meeting of the Association des médecins psychiatres du Québec (5). It illustrates a historical (and, to some extent, persistent) difference between Anglo-Saxon and European (in particular, Latin) psychiatry. On this side of the Atlantic, McMaster University and the Cochrane collaboration have played a key role in the development of EBM and evidence-based psychiatry (EBP). Of course, the frontier lies between the Canada-US continuum on one hand and Mexico, Central, and South America on the other. The Anglo-Saxon tradition is more attracted to objective and quantified reality, while the philosophy underlying Latin practice is somewhat more focused on subjective reality. But, in the past couple of decades, subjectivity (under the names of consciousness studies, cognitive neuroscience, and affective neuroscience) has become an object of science itself, and Claude Bernard’s statement, “L’art est je, la science est nous,” may soon become obsolete. The mind–brain problem, for millennia the province of metaphysics, has become a legitimate field of scientific inquiry, with Nobel prizes in physiology and medicine (awarded to Eccles, Edelman, Crick, and Kandel) focusing on the emergence of subjectivity out of neurons. The frontier between objective and subjective categories is blurring quickly. A Canadian research network on this topic has recently been proposed (6) and has received support from the Canadian Institutes of Health Research.
Philosophers of mind and neuroscientists are split between the emergentist materialists (for example, Patricia Churchland and Jean-Pierre Changeux) who expect to be able to reduce all the phenomenal aspects of subjectivity (and, in particular, all human emotions) to neuronal functioning—thus making them eventually objectifiable—and those who do not share this belief (for example, Eccles, Penfield, and Ricoeur).
The classic tools of EBM are randomized clinical trials (RCTs) and the metaanalyses which build on them to draw general conclusions. A good example is Project MATCH (Matching Alcoholism Treatment to Client Heterogeneity), certainly the largest and most ambitious RCT to evaluate psychosocial or psychotherapeutic approaches in the treatment of alcoholism (7). Costing more than USD 30 million over 5 years, this meticulous, multicentric study compared 3 treatment methods. Its working hypothesis was that individual subjects’ characteristics could be matched preferentially with a particular approach.
The quality of the research protocol, the planning by the best researchers in the field, the strict precautions to select and train the therapists and ensure that they followed precisely the manualized therapies, the perfectionism of initial assessments, and the high quality of follow-up and outcome measures have been such that results proved spectacularly superior to expectations regarding the usual outcome with these well-established techniques. The attrition rate was minimal, and the matching hypotheses were essentially not confirmed. In spite of all their scientific qualities, however, the radical requirements of the best RCTs may make them irrelevant from the public health viewpoint, and often in clinical practice, because of the many exclusion criteria.
Beyond this ideological dilemma, EBP may have to broaden its definition of “evidence”—even though it already accepts qualitative as well as quantitative research methods: in the RCT, “real” evidence is a process in which an active molecule interacts with a receptor. It is compared with placebo, the symbol of an elementary psychotherapeutic (that is, subjective) interaction. Brain-imaging studies provide objective evidence that the human organism is also very susceptible to such contextual variables as interpersonal relationships; now, cognitive neurosciences have become able to focus on certain systems in the brain (for example, those dealing with affective states such as anxiety, pain, or depression) that are much more responsive to interpersonal influences than are others. Placebos have been found to relieve depression and pain better than they do most illnesses, and genuine neurobiology (8) cannot anymore ignore interpersonal and historical influences on mind–brain functioning. It may explore the placebo effect in a richer perspective than is offered by the traditional placebo-controlled studies. The preoccupation of the clinician should be: “How can I maximize the placebo–response component that we are producing anyway in the routine care of depression?” Conversely, the clinician–scientist’s priority will remain: “How can I minimize the placebo response, so that I can demonstrate beyond any doubt the efficacy of the active medication?”
Nevertheless, in spite of some clear divergences, Canadian and British contributors to the present issue agree on one essential lesson in their reviews: in spite of meritorius effort, clinical practice guidelines are still often of dubious quality, and the jury is still out on much of the evidence upon which we all hope to build scientific psychiatry.
1. Lesage A,
Stip E, Grunberg F. What’s up doc ? (2001) Le contexte, les limites et les
enjeux pour les cliniciens de la médecine basée sur des données probantes
(Evidence-Based Medicine). Can J Psychiatry 2001;46:396–402.