There are a lot of theoretical arguments in favor of science-based and evidence-based medicine and against the alternatives of opinion-based, tradition-based, or faith-based medicine, and I have discussed many of these (for example, Medical Cognition, Spiraling Empiricism, the Dunning-Kruger Effect, and a host of reasons why clinical judgment is unreliable). However, the bottom line that I always come back to is that evidence-based medicine works! This is evident from the simple and obvious story of medical history, which illustrates the dramatic and unprecedented improvement in the length and quality of human life that science and its application to sanitation, nutrition, and medicine have brought us. But it is also evident in a more quantifiable way in a new study which found that the adoption of consistent, evidence-based standards of care for hospital patients with heart disease resulted in a marked improvement in both short and long-term survival.
The study was conducted in Sweden and involved a national database that covers almost all hospitals in the country. It included records for over 67,000 patients treated over a 12-year period for a particular kind of abnormal heart rhythm associated with a heart attacked. The study demonstrated that specific evidence-based practices, including surgical procedures and the use of new medications, gradually increased over the course of the study. Along with this, in-hospital complications decreased, and survival improved markedly at 30 days, 1 year, and 12 years after the heart attack. Looked at another way, this translates into an average of almost 3 years longer survival for a patient treated at the end of the study, with more evidence-based methods, than at the beginning of the study.
For all the confidence we have in our judgment and experience, and those of our mentors, and all the optimistic promises made for ideas justified by tradition and anecdote, the bottom line is that the slow, laborious, rigorous process of basing our treatments on ever-changing scientific evidence simply works better.