Limits of Evidence Based Medicine revisited

In 1996 D.L. Sackett wrote in the BMJ « Evidence based medicine : what it is and what it isn’t » and the characteristics and limits of EBM remain of current interest. We will touch upon some highlights.

EBM aids clinical decision making based on medical evidence. And there is no need to question the thesis that medical practice should be based on the best available evidence, if indeed such evidence is available. However the implementation of the principles of EBM is complex.

If EBM introduces science in the art of medicine, the question is : what can we consider as best evidence? What is the role of the clinical experience in the evidence? Can EBM be restricted to randomized controlled trials? The study of the presupposition of randomized controlled trials shows that RCTs are frequently overvalued in comparison with cases studies.

But often there are few or no data, or they fall in ambiguous “grey zones”, and the evidence hierarchy of the data can be questioned, especially since pre-eminence is given to randomized controlled trials. The restriction of EBM to randomized trials and meta-analyses based on published data seems to us adogmatic reductionist approach to deal with the real uncertainty of the clinical setting that EBM intends to remedy. A review on a treatment or a procedure may conclude that there is “no evidence of benefit” because no RCT has been conducted on that treatment or procedure.

EBM gives more importance to the care of populations than to the care of individuals. The statistical results are focused on the results of large groups of patients.

There is a big gap between clinical research and individuality of patients. Often the patients included in RCTs differ from the patients seen in everyday practice.

Outcomes of studies (RCT) might be biased as their sponsors designed them to obtain positive results. Many randomised trials exclude large, relevant patient populations, such as children, pregnant women, patients with comorbidities and the elderly. It is not unusual that different trials and meta-analyses lead to contradictory results.

Moreover, the decision making models based on purely quantitative, statistical approach of the “best evidence” have shown their limits in the actual clinical

setting dealing with individual patients, as well as in a broader epistemological context. For example, modern surgery integrates all dimensions of patient care, not only particular surgical procedures. Very few high quality RCTs of surgical treatments exist. Current surgery and surgery of the future mostly improve with new technologies. Surgical techniques evolve continuously, and faster than the time needed to conduct a trial.

A surgeon works with a team, and every member of the team must be skilled, and RCTs does not study or control the clinical environment.

There is also a gap between empirical evidence and clinical practice. To take the best clinical decision physicians and patients should also take into account empirical and experimental evidence, physiologic principles, and what the patient and the physician consider to be the best. Clinical decision relies on exploitation of best evidence and the phenomenological attention focused on the patient and his environment.

The question which must be considered is the transfer of knowledge to clinical medicine. Medical art asks the fundamental question of transfer and integration of knowledge. Clinical decision relies on the best clinical evidence and on contextual data that are not countable.

And last but not least, is patient choice based on clinical evidence, which is a key element of the EBM paradigm, real or a mere illusion?

Evidence-based medicine is a useful tool but it is limited and non exclusive.