I often ask my students, “Will two well-done guidelines that are free of commercial bias and that effectively use the same excellent literature come to the same conclusion?” The first thought, especially for those of us who support evidence-based medicine, is “Yes,” or perhaps “I hope so.” The answer, however, could be “Maybe not.” This is because there is a lot more to guideline development than evidence, as illustrated by two recent guidelines related to cardiovascular risk and children with attention-deficit/hyperactivity disorder (ADHD): an American Heart Association (AHA) guideline,¹ which argued that it is reasonable to obtain routine electrocardiography (ECG) before starting stimulant medication, and the subsequent response from the American Academy of Pediatrics (AAP),² which does not recommend routine ECG in most children before starting ADHD treatment. The Practice Guidelines on these topics can be found in this issue of American Family Physician.^3,4

How can two reputable groups apply evidence-based methodologies to the same data and end up with different recommendations? Is there a basic flaw in evidence-based medicine? No, I don't think that is the problem. In some cases, there may not be enough evidence, but the problem here is the guideline process itself—too often the process remains implicit rather than explicit.

Most guidelines provide an evidence profile that makes explicit statements on the quality of the evidence and the strength of the recommendation. However, two persons looking at the same information often draw different conclusions.

Authors of an article about developing evidence-based guidelines propose to expand the evidence profile.⁵ In brief, they suggest that the evidence profile should explicitly discuss: (1) the quality of the evidence; (2) benefits of implementing the recommendation; (3) harms of implementing the recommendation; (4) costs; (5) the benefit-harm balance assessment; (6) the value judgments the panel used in evaluating this balance; (6) the role of patient preferences; and (7) the strength of recommendation.

Using this evidence profile model as a framework, one can easily see how two different groups looking at the same evidence could come to different recommendations. (In addition, there is some indication that recent American College of Cardiology/AHA guidelines are largely based on lower-quality evidence and expert opinion.⁶) For example, perhaps the AHA did not consider that getting an ECG may have associated harms. However, if it did, it might have determined that any benefit would outweigh the harms and costs. This is a common problem with guidelines, and it can be at least partially addressed by assuring a breadth of perspectives on the guideline panel.

So what are we as family physicians to do? Take the AHA's position, because that group has the content experts? Or the AAP's position, because they also include primary care physicians? I think it is dangerous to blindly do either. Guidelines are not designed to dictate, but to guide. When faced with conflicting guidelines, physicians must decide which to follow. Factors that might affect this decision include the endorsement of a guideline or other clinical policies by a national medical organization. The American Academy of Family Physicians currently does not have a position on these guidelines. We could also look for commentaries about these guidelines that can present arguments to help us decide whether one makes more sense than the other.⁷ Physicians might turn to the medical community in which they work to see which clinical policies it follows.

In the case of the AHA and AAP guidelines, one group states that routine ECG is reasonable, and another states that it is not needed. Although these two statements are not directly contradictory, they are far from being in agreement. I am frustrated by this disagreement, but I think I understand the reasons why. Without evidence, we do not know which is the correct approach, but we are forced to decide what to do in practice. My thought is that in the absence of evidence, the prudent approach is often the less invasive one. Although the AHA approach may be reasonable, the AAP position of not obtaining a routine ECG before prescribing stimulant medication for ADHD seems, to me, the more prudent.