All Opinions Are Not Equal
To the Editor:
Gibbons et al1 detail the “noble” and Herculean task of developing American College of Cardiology/American Heart Association cardiovascular guidelines. Recommendations are based on either experimental results or opinion, with the latter being considered of uniformly lower evidentiary importance. The contribution of opinion to the guidelines is substantial in that opinion supports more than half of the current recommendations and appears to contribute continuity and cohesiveness to recommendations that would otherwise be fragmented in large part as a result of insufficient experimental data. On closer reading, it becomes apparent that the treatment of all nonexperimental support as “opinion” masks at least 2 different, distinct, and incommensurable, types of opinion evidence. A low level of evidentiary importance may be quite appropriate for the many opinions that deal with issues that are open to experimental testing when this testing has not been done for a variety of reasons ranging from a lack of interest in funding to ethical issues regarding controls. However, other assignments of opinion to level C evidence ignores that our search for scientific truth takes places within a framework of other truths and assumptions. For example, the need for a clinical history and tests that are intrinsic to the definition of disease states often is not an opinion that is subject to scientific testing but rather is a presupposition of our scientific testing. Moreover, our work proceeds within moral and legal constraints. Moral imperatives, such as “Do no harm,” form the basic fabric of our profession and have predated and will outlive much of our current experimental truths. Opinion evidence often reflects moral imperatives, including not treating when there is no indication, when there is a significant contraindication, when risk outweighs benefit, when the patient refuses treatment, or when treatment cannot be effective. Moreover, no matter how therapeutically efficacious an illicit drug is, it must receive a class III not a class I recommendation. These types of truths and assumptions are arguably of equal weight to level A or B evidence. Hence, in important respects, the current hierarchy of evidence blurs scientific assumptions that could plausibly be subject to testing, with other truths that are not, and cannot be, evidence-based. It is a logical error to group these distinct types of “opinions” as one and assign them the same level of evidentiary import.
Gibbons RJ, Smith S, Antman E. American College of Cardiology/American Heart Association Clinical Practice Guidelines, part I: where do they come from? Circulation. 2003; 107: 2979–2986.
We appreciate Dr Kessler’s interest in our article describing the American College of Cardiology/American Heart Association Clinical Practice Guideline effort. We agree with him that many of the guideline recommendations are based on expert opinion rather than evidence and that these recommendations contribute to the clinical utility of the guidelines. At the same time, these recommendations also serve to identify the many areas where evidence is lacking.
Many major medical and governmental organizations have published methodology for the development of practice guidelines1–5 that emphasizes the importance of scientific evidence. The ranking of evidence in ACC/AHA Guidelines is consistent with these published standards.
Dr Kessler suggests that some types of expert opinion should be considered equal to randomized trials, and that these include “moral imperatives including not treating when there is no indication, when there is a significant contraindication, when risk outweighs benefit…” We disagree. Such current “moral imperatives” are subject to rapid change as new scientific evidence is available. For example, in the 1970s, there was no indication for fibrinolytic therapy in acute myocardial infarction, acute myocardial infarction was a contraindication to the use of sublingual nitroglycerin, and the risk of coronary angiography during acute myocardial infarction was felt to far outweigh the benefit. All of these “moral imperatives” have now changed in response to new scientific evidence. The transient nature of current opinion explains the focus on scientific evidence in the development of clinical practice guidelines and the hierarchy of grading evidence that gives expert opinion the lowest level of evidence (C). This hierarchy is not a “logical error,” as Dr Kessler contends, but rather a logical effort to acknowledge the current gaps in the available scientific evidence.
Committee to Advise the Public Health Service on Clinical Practice Guidelines, Institute of Medicine, Field MJ, Lohr KN, eds. Clinical Practice Guidelines: Directions of a New Program. Washington, DC: National Academy Press; 1990.
Eddy DM. A Manual for Assessing Health Practices and Designing Practice Polices: The Explicit Approach. Philadelphia, PA: American College of Physicians; 1992.
American Medical Association, Office of Quality Assurance. Attributes to Guide the Development and Evaluation of Practice Parameters. Chicago, IL: American Medical Association; 1990.
Canadian Medical Association. Quality of Care Program: The Guidelines for Canadian Clinical Practice Guidelines. Ottawa, Ontario: Canadian Medical Association; 1993.
Woolf SH. Manual for Clinical Practice Guideline Development. Rockville, MD: Agency for Health Care Policy and Research; 1991.AHCPR publication 91–0007.