Evidence‐based medicine: Progress but not a final solution

Convincing evidence of the beneficial effects of a treatment used to be provided by the consistency with which it relieved symptoms and abnormal findings andor prolonged survival.The beneficial effects of diuretics for relieving fluid accumulation in patients with cardiac failure and those of bypass surgery or angioplasty for improving effort tolerance in patients with flow-limiting coronary stenoses were so obvious and consistent that there had been no need for large multicenter trials.However, the benefit of some treatments may not be obvious, and therefore the consistency of their effect may be uncertain: this may occur when a treatment has only a moderate effect (or additional effect) on the pathogenetic mechanisms of the disease, or when the underlying disease causes only mild symptoms and abnormalities and is seldom associated with adverse events.In the past, the choice of such treatments was left to a sort of clinical freedom, or even worse, of clinical anarchy: it was based on personal experience or, worse, on opinions.The advent of randomized clinical trials represented a major, rational step forward because trials provide a statistically based quantitative comparison of average benefits and side effects of the treatments tested.Nowadays, multicenter, randomized trials are indeed the acid test for any therapeutic hypothesis derived from experimental findings or from pilot studies: they have become the basis of "evidence-based medicine."This concept is now gaining a wide acceptance because it provides useful standards of care and protects from malpractice allegations.The design of trials, their statistical analysis and meta-analysis have become so sophisticated that major journals were compelled to recruit statisticians onto their editorial boards.