How to incorporate clinical experience into evidence-based medicine.

Evidence-based medicine (EBM) represents a paradigm change in the way in which medicine is practiced. On one level, evidence-based medicine is a formalism for integrating “research evidence with clinical expertise and patient values.”5 It is not intended to supplant clinical experience, but rather to augment and strengthen it. There has been a concern voiced within the neurosurgical community that EBM will reduce patient care to a set of protocols and lead to a cookbook approach to patient care. There is a temptation to allow EBM and its emphasis on the literature to cut discussion short rather than to broaden clinical considerations. Not to recognize the fundamental importance of clinical experience, however, is a clear distortion of the original intent of EBM. The passions that are inflamed by the EBM movement are fanned by questions of what it means to know something. Contemporary clinical medicine is an exercise in reasoning with uncertainty. The data that we collect are imprecise and often subjective. It is well accepted that a diagnosis is a prerequisite for effective treatment. Yet at the same time we use the concept of a differential diagnosis to acknowledge the fact that our final determination may be incorrect and that other etiologies are possible. In the end, however, we have a need to know that the treatment plan that we are implementing is correct, and this desire drives us to assert more certainty than we know to be the case. To be wrong is unacceptable. In this context, it is helpful to point out that a paradigm of absolute reliance on the literature is no more likely to lead to optimal management decisions than a blind adherence to clinical experience alone. Literature reviews can be conducted using a meaningful, reliable, and repeatable methodology. The fundamental question, however, is whether an article or body of work is applicable to a particular clinical question. Often the answer is unclear. Some decisions are driven more by the efficacy of treatment, whereas others are more closely tied to the risk of side effects. The first step in implementing EBM is formulating an appropriate clinical question. This is where much of the art of EBM resides and also where fundamental and important assumptions creep into the process. The contention that EBM is an objective process and without bias should be debunked once and for all. Just as it is possible for a decision to be steeped too heavily in the clinical experience of one expert, it is also possible for a decision to rely excessively on the medical literature without accommodating the unique particulars of the individual case. How then are we to proceed? Stephen Haines has observed that the salient insight of the EBM paradigm is that the quality of evidence is of vital importance when applying conclusions from the literature to solve clinical problems.3 It is well accepted in the neurosurgical community that 100 case reports do not carry the same weight as one well-implemented, randomized, controlled trial (RCT). Central to EBM is a system for categorizing the importance of the data (Table 10.1). When well-designed and well-executed RCT studies that speak to the clinical problem in question exist, they obviously bring significant clarity to clinical decision making. Unfortunately, the likelihood of such an occurrence on a typical neurosurgical service is rare. There are few questions in neurosurgery, as in the other medical specialties, for which RCTs have been performed. There are many obstacles to the implementation of an RCT. In addition, there are some situations that make the institution of an RCT unnecessary or ill advised.6 Sackett5 identified a clinical situation called all or none in which before the institution of a particular treatment, a disease is universally fatal, whereas with the treatment, some patients survive.7 He has argued that an RCT is unethical in the study of an all-or-none disease. Another example of an all-or-none situation occurs when some patients die of the disease before treatment, but all survive after treatment is initiated. In other less stark situations in which there is a lack of clinical equipoise, it is unethical to randomize patients. Fortunately, RCTs are not a requirement to practice sound EBM that incorporates clinical experience. As mentioned above, one of the first steps in EBM is determining whether the evidence is applicable to the current clinical situation.7 It is important to ask whether the patient Copyright © 2009 by The Congress of Neurological Surgeons 0148-703/09/5601-0054