Thomas Huxley characterised “the great tragedy of Science” as “the slaying of a beautiful hypothesis by an ugly fact.” Unfortunately, when medical hypotheses are disproven, they act more like zombies than corpses, revived by the sorcerers of Mammon, aided and abetted by the inertia of medical practice.

Recent examples in diverse areas of clinical practice will likely suffer the problem of perseverance of beautiful but flawed hypotheses, so we are raising the alarm now. Three large trials in 2008 showed that intensive control of type 2 diabetes mellitus lacks benefits for patients and increases adverse effects1-3 (including one in this issue of Evidence-Based Medicine4), and 2 trials showed that the self-monitoring of blood sugar in type 2 diabetes is not cost-effective5 and is associated with depression.6 Two trials have documented the lack of benefits of antiviral agents for Bell palsy (while confirming the benefits of corticosteroids).7 8 Many trials and meta-analyses have confirmed and reconfirmed the absence of benefits and presence of harmful effects of antioxidants for the prevention of cancer9 10 and cardiovascular disease.11

For each of these “slayings,” vested interests will undoubtedly try to make us forget that the justification for their promotions has been gored. For type 2 diabetes, for example, achieving a haemoglobin (HB) A_1c ⩽7% means selling more drugs and more glucose monitor strips, despite both being detriments to the health and happiness of patients. It is unlikely that the advertising and promotion will desist without legislation.

Unfortunately, even if the promotion of disgraced medical interventions were to desist, prescription by licensed practitioners would persist for decades. This sad fact of human behaviour is nothing new. Samuel Johnson, the 18th century poet and critic, observed that “The chains of habit are too weak to be felt until they are too strong to be broken.” At the turn of the 20^th century, renowned physicist Max Planck remarked: “A new scientific truth does not triumph by convincing its opponents and making them see the light, but rather because its opponents eventually die and a new generation grows up that is familiar with it.”

In more modern times, social psychologists have empirically studied people’s tendency to cling to their cherished beliefs, even in the face of a mountain of disconfirming evidence, known formally as the confirmation bias. Once a belief is established, people will actively seek out information that supports their belief. If confronted with mixed evidence, people will give greater weight to the information that supports their belief while discounting contradictory information. For instance, in 1 experiment, groups of people with pro- and anti-capital punishment beliefs were presented with an empirical report that either concluded that capital punishment was effective or ineffective. When the conclusion did not support their beliefs, both those for and against capital punishment were hasty to point out methodological flaws of the research design and found the results unconvincing. When the conclusion supported their beliefs, however, people praised the research for its sound design.12

Can we do no better than this in medicine? Apparently not. A recent study documents the persistence of outmoded claims in the medical literature.13 Quality of care studies show that our profession is always behind the evidence curve, with adoption of validated new procedures that is <50% many years after the evidence has become clear. Worse still, studies show that didactic instruction, still the predominant mode of continuing education, has no discernable effect on clinical practice, and practitioners are quite resistant to seemingly vigorous interventions such as audit and feedback, preceptorships, patient-mediated interventions, educational influentials, and computerised decision support.14 15

What does it take to put an ugly fact through the heart of a beautiful hypothesis? Clearly facts by themselves are not enough or at least not fast enough. Individually, we need to do what we can in our own practices. For type 2 diabetes, the recent trials mean backing off trying to help patients reach the “goal” of HbA_1c levels <7% and, if they are not on insulin or prone to hypoglycaemia, not advising glucose self-monitoring unless the patient requests it. It is not clear exactly what the glycaemic goal should be, but the patients in the control groups of the new trials, whose HbA_1c averaged between 7.3% and 8.4%, did at least as well as those with mean HbA_1c <7%, so advising patients to aim for levels of 7.5% seems defensible.

But imprecations to individual practitioners to do better have never worked very well, and it is clear that we do not know enough about how to maintain professional competence over time. Research funding agencies have realised for some time that the practical return on investment for health research is far short of societal expectations. Recently, stimulated by the “show us the evidence” questions of policy makers, funders have begun to put resources into “knowledge translation (KT) research” at 2 intersections, from laboratory to clinical research (KT1) and from clinical research to health care (KT2).16 This funding recognises our insufficient knowledge of how to effectively and efficiently translate research from one level to the next, and that there are legitimate and important research questions about research translation. Ironically, the attitude of some members of the research community, that we simply need to try harder at broadcasting our research, and that translation research is not needed let alone legitimate, has been slow to change. However, it is changing and a cohort of KT researchers is emerging.

When will we be able to slay beautiful hypotheses more efficiently? No one knows, but at least KT research is becoming fashionable and is beginning to address the right questions. Meanwhile, if the new evidence differs from the advice you have been giving to your patients about antiviral agents for Bell palsy, intensive care for type 2 diabetes, or antioxidants for preventing cancer or cardiovascular disease, it is time to consider revising your practice. According to Pathman et al,17 the logical steps for this include becoming aware of the evidence (awareness), then agreeing with it in principle (agreement), deciding it is appropriate and feasible to use in your own practice (adoption), and finally succeeding in following it at appropriate times (adherence), a model that is consistent with guideline implementation for hypertension management in primary care settings.18