The limits of evidence-based medicine
By David Annis
Article ID: 1244
For many years, doctors have been free to practice medicine in almost any way that they saw fit. They could use drugs to treat conditions for which they were not approved and use any procedure on any patient from whom they obtained informed consent. At times this led to doctors treating patients in a manner not supported by scientific research.
Recently, there has been a vigorous movement to get doctors to practice “evidence-based medicine”. This movement advocates treating every disease based on the available scientific evidence about treatment efficacy and risk. I have been a strong proponent of this; after all, how can I oppose those who espouse belief based merely on faith and then not support evidence-based medicine?
However, I’ve lately been forced to rethink my position on evidence-based medicine. It sounds good initially, but there are limitations to its real-world applications:
- The subjects in drug trials and other studies are seldom an exact match for the patients that a doctor is treating. The trial may have been done using patients in a different age range and physical condition. Most trials purposely exclude people with medical issues other than the one being studied. A treatment that was well tolerated in a patient population with a median age of 35 may not be appropriate for a 60 year old with renal insufficiency.
- Diagnostic categories are sometimes not exact. For example, a patient may have some symptoms that are consistent with ADD and others that are consistent with Autistic Spectrum Disorder. A doctor – forced by the insurance company to choose – may diagnose ADD, but still wish to treat for symptoms of Autism Spectrum Disorder.
- All evidence is not equally good. Some studies are better designed than others, and sample sizes and levels of statistical significance vary. There have been attempts to assign a level to each piece of evidence to indicate its quality, but there are still differences within a level. And levels can sometimes be misleading based on the rules used to classify the evidence.
- Even if the quality of the evidence is classified properly, evidence can be contradictory. How many weaker studies are needed to overcome the results of a stronger study? The answer is subjective.
- The process of developing a set of evidence-based guidelines is not an entirely impartial and objective process. Insurance companies try to influence the process in a way that keeps costs low. Drug companies try to influence the process in a way that ensures their products are included in the standard protocol. Those that evaluate the scientific evidence may have their own biases affecting the outcome.
A good doctor looks at the specific individual in front of him, uses his knowledge of how the body works, the patient’s specific constellation of diseases, the patient’s physical and mental state, and decides on the course of treatment that is best for that individual. Attempting to reduce that clinical judgment to a flow chart and checklist does a disservice to both the patient and the physician.
I am certainly not advocating allowing doctors to treat pain with voodoo dolls and crystal power, but I do believe that we are in danger of allowing the pendulum to swing too far. Physicians are highly educated professionals, dealing with highly complex, poorly understood systems, including the human mind and body, disease causing organisms, and the natural fauna that exists in and on all of us. They also have to process a continual stream of new information and new discoveries. Reducing clinical care to a set of checklists and decision trees is a disservice to us all.