One day a residency program decided to put its evidence-based medicine (EBM) curriculum to good use. A group of faculty and residents conducted a thorough review of the evidence regarding liquid-based cytoloutilizegy vs conventional Pap smears. They identified the key national recommendations and reviewed the supporting evidence behind each recommendation, tracing back to the individual studies themselves.1-3
Based on the review, the group concluded that there was insufficient evidence to recommend one method of screening over another, but that there were situations in which one method might be preferred. They presented the evidence and their conclusions to the majority of faculty and residents at grand rounds. Following the presentation, the larger group discussed the relative merits of each screening method and decided the elements of evidence that supported the liquid test were more relevant to the practice than the conventional Pap smear. As a result, a decision was made by the group to stop carrying supplies for the conventional Pap smear. While the decision seemed reasonable on the level of an individual practitioner, several faculty and residents were unhappy with the “evidence-based” decision.
KAP theory and EBM
KAP theory identifies Knowledge, Attitudes, and Practice beliefs as key elements that drive healthcare providers’ decisions about medical care. In a sense EBM represents knowledge.4 There is a collective body of medical knowledge in the form of research, which represents “the evidence.” And there is what the healthcare provider himself “knows.” A major purpose of healthcare recommendations, point of care information systems, and best practice guidelines is to help the healthcare provider’s individual medical knowledge reflect the collective body of evidence.
For the purposes of this example, evidence will be considered absolute, inadequate, conditional, or relative. Absolute evidence occurs when there is clearly a correct answer. For example, the net benefits of aspirin for the treatment of myocardial infarction are clear. However, for most topics the evidence is not absolute; rather, it is inconclusive.5 The evidence may be inconclusive because it is inadequate—eg, insufficient research, conflicting studies, or research on peripheral topics. As an example, studies have demonstrated that aspirin decreases colorectal polyps, which may or may not be peripheral to the question of whether aspirin prevents colorectal cancer.
The evidence can also be conditional, meaning that in some defined instances the net benefit is clear. However, extending this net benefit beyond these instances is less clear. For example, patients at high risk for cardiovascular disease have a clear net benefit in taking aspirin for myocardial infarction prevention. Finally, the evidence may be relative, with a balance of known benefits and known risks.6 Using the aspirin example for cardiovascular disease prevention, patients at moderate risk receive benefit from aspirin in preventing myocardial infarction but at a risk cost of increased bleeding.
When the evidence is inconclusive, the second and third aspects of KAP theory—attitudes and practice beliefs—become very important. Healthcare providers and patients may arrive at different conclusions based on different viewpoints. On an individual level, healthcare practitioners use tools such as shared decision-making and patient-centered care to reach decisions.6,7 However, inconclusive evidence provides a unique challenge when trying to develop local, regional, or national standards.
Evidence heresy
EBM frequently has negative connotations. In a room full of healthcare providers, some will believe that EBM should revolutionize the practice of medicine,8 and some that EBM has limited utility.9 How does this happen? The above scenario serves as a useful example, highlighting 3 misuses of the term “evidence” that frequently give EBM a bad name.
First, inconclusive evidence should not be stated in absolute terms; rather, it is more helpful to explicitly state what we know and the limits of what we know. Shaughnessy and Slawson wrote, “Absolute certainty is absolutely impossible, and we do not have to wait for that, of course.”10 This reflects the paucity of topics with certain evidence and highlights the need for clinicians to act on the available information. Every clinician necessarily utilizes this skill on a daily basis. The clinician has to become an Information Master11 and know not only the end result of what the evidence indicates but also the facts supporting the end results and how those facts apply to the care of an unique individual.12 However, taking this a step further and stating that one answer or option is absolutely correct in all cases ventures into dangerous ground. During the residency’s discussion of cervical cancer screening tests, the group recognized the merits of both options verbally but, the act of removing all conventional Pap smear supplies implied the nonverbal judgment that liquid-based technology was an absolute correct answer.