July 2017
On Medical Reversal

There is an old joke in medicine that goes: on graduation day the dean of the medical school tells the new doctors that half of what they learned will turn out to be wrong – we just don’t know which half. Unfortunately, it turns out to be more than half. In 2005 Dr. John Ioannidies wrote a seminal article entitled “Why Most Published Research Findings are False.” Since then there has been a focus in the Evidence Based Medicine (EBM) movement on improving how studies are done and preventing reversals in medicine, but there is much resistance to be overcome. Why is this important? Bad medicine hurts patients; even more importantly in the long run, people lose faith in science and medicine.

The public is whipsawed by medical reversals such that ordinary people sense that something is fundamentally wrong with our system. First they are told that a high fat diet is bad and causes obesity, and later that that was wrong and that it is really a high carbohydrate diet that causes obesity. Whoops! Never is there an admission that nutritional study is unreliable because the human diet is complex and cannot be controlled for the sake of study. Butter is bad for you - use margarine instead! No, no, what we meant to say is that margarine is bad and you should use butter instead. Whoops! How many patients had to die on the nonnarcotic pain medication Vioxx or the antibiotic Trovan before they were withdrawn from the market by the Federal Drug Administration (FDA)? This tale of woe just goes on and on, reinforced in the mind of the public by legal advertising: Call 1-800-BAD-DRUG.

Dr. David Sackett is considered to be the father of the EBM movement that studies medical practice scientifically to determine what really works. Nevertheless, much of what we do as doctors has either not been studied or cannot be studied for practical or ethical reasons. In those cases we continue to rely on tradition, anecdote, and direct experience. Sackett has identified more than 100 forms of bias that can undermine findings and make moot all the fancy statistical footwork that accompanies the study. Bias can be very subtle. For example, drug companies rely on local doctors’ offices to test new drugs. The doctors who engage in this research must hire a certified clinical research registered nurse and restructure their offices at considerable expense, hoping for drug company research money. They will get the sought after contracts and keep getting them as long as they play along. However, if they report too many serious side effects then their contracts dry up and their research business fails. Bias in drug company and medical device manufacturer research is the greatest current obstacle to achieving honest results in medical research.

(To be continued…)