Evidence-Based Medicine

Evidence-based medicine refers to an approach to the teaching and practice of medicine that developed in the late 1980s and early 1990s. The approach emphasizes that medicine should be practiced according to evidence that can be assayed on the basis of its quality. The idea is that from among competing claims that physicians might face concerning various courses of action they might take, they can choose a course from evidence that has been assessed rigorously and presented systematically.

An influential aspect of the evidence-based medicine approach has been the Cochrane corroboration which is a collaboration of physicians and scientists who undertake detailed, systematic reviews of the evidence surrounding given courses of action. They attempt to gather all relevant information, including research that was conducted to a high standard but its results were not published. (Failure to publish the results of clinical trials appears to be common in the pharmaceutical industry, where there is no incentive and perhaps even a disincentive, to publishing "negative" results, i.e., trials that did not show a given drug to be beneficial.) By assaying all of the information, and prioritizing it by the rigor with which the data were collected, the hope is that a clear picture on which treatment is best will emerge.

For example, consider a patient who is diagnosed with pneumonia, but in whom its cause is not clear (e.g., is it caused by a bacterium? which one?). While almost all physicians will have been taught how to treat pneumonia, in an ideal situation they could rapidly consult an electronic database in which all rigorous studies of treating pneumonia have been reviewed. At present, the-highest scientific standard is generally the randomized, blind, controlled trial. In such a trial, patients are assigned to drug A or to drug B, where drug B is either the usual standard (as would be the case in a pneumonia trial) or a placebo (as would be the case in a condition for which no drug has a proven benefit compared with a placebo). Since the patient is randomly assigned to either the new interaction (drug A) or a control group (B), the trial is randomized and controlled. It is also important that the trial be blind, usually double-blind (i.e., neither the patient nor the treating physician is aware of which patient is in which group), so that wishful thinking cannot easily play a part in the decision whether the new drug works better than the existing standard).

For many of the early decades of scientific medicine, the existing standard appeared to be the best available, but as the number of well conducted trials increased, there was a need to aggregate them and summarize the general experience. In consequence. the review article, written by an expert who might have a particular point of view and who might selectively choose evidence, gave way to systematic review in which the evidence is collected in a standard fashion (usually beginning with an electronic literature search) and summarized in a standard way (with procedures known as meta-analyses).

Though no one would argue that doctors should pursue treatments that do not work, there is controversy over how to understand information produced from evidence-based systematic reviews. The controversy is sometimes technical in origin; for example, there are disputes over which is the best way to carry out particular studies or whether it is more important to understand the quality of a given clinical trial than its result, or whether, when the trial experience is aggregated, all trials should count equally. There is also some skepticism that, even using multiple observers, results are still subject to the bias of the reviewers, and thus systematic reviews by different groups can give rise to substantially different conclusions and guidelines.

While these arguments are substantial, there are other areas of controversy that relate to how a given body of information about the average experience of many patients can be translated into a course of action for a given patient. Should the evidence completely determine the course of action, to the extent that another course might be disallowed, or not paid for, or subject to review by a third party? Or should the evidence about what generally works be factored in as one among other important considerations about what to do for a given patient? At this stage, the answer is not clear, and there are thoughtful proponents of each course of action. Nevertheless, the evidence-based medicine movement in seeking to improve scientific standards of clinical inquiry, and in seeking to reduce the arbitrariness of how data are presented, has done much to improve the quality of evidence available to clinicians.