Nearly half of the current clinical practice recommendations issued by the American College of Cardiology and the American Heart Association are based on expert opinion, case studies, or standards of care, and hence are not evidence based.
This may or may not be a problem, but it highlights the need for more studies to produce the data to fully substantiate how cardiovascular disease should be managed.
The American College of Cardiology and the American Heart Association “fully support the idea that as much as possible, we should have an evidence base, but there are many clinical situations where the studies have just not been done,” said Dr. Sidney C. Smith Jr., professor of medicine and director of the center for cardiovascular science and medicine at the University of North Carolina at Chapel Hill. And in some cases it does not make sense to test standard practice in a controlled study, such as running an ECG on patients with chest pain, he noted.
Dr. Smith disagreed with the notion that practice guidelines should be produced only when study results back them up. “There are situations when providers, patients, and payers need recommendations; it's important that we get the best opinion possible and indicate that it is expert opinion and not based on the results of a randomized, controlled trial.
“Evidence-based medicine has tremendous promise for our patients, and we've made tremendous progress in the past decade to develop recommendations based on evidence, but we have much more work to do. It's a good time to invest” in better medicine, said Dr. Smith, who is also a former chief science officer for the AHA and currently chairs the ACC/AHA Task Force on Practice Guidelines.
Dr. Smith was the initiator of and coauthor of a recently reported study that quantified the level of evidence behind all 2,711 practice recommendations contained in the 16 current guidelines promoted by the ACC and AHA joint program (JAMA 2009;301:831-41). Among these recommendations 11% were backed by level A evidence, defined as evidence coming from multiple randomized trials or meta-analyses; 39% were rated as having level B evidence, defined as evidence from a single randomized trial or from nonrandomized studies; and 48% were derived from level C evidence, meaning expert opinion, case studies, or standards of care (total is less than 100% because of rounding).
But not all experts who deal with crafting clinical practice recommendations take as benign a view of basing them on level C evidence.
“I think expert opinion is quite misleading” when used as the basis for a practice recommendation, said Dr. Diana B. Petitti, professor of biomedical informatics at Arizona State University in Phoenix and vice chair of the U.S. Preventive Services Task Force (USPSTF), a panel organized by the federal Agency for Healthcare Research and Quality to formulate practice recommendations for clinical preventive services.
“Expert opinions imply that there is something that the experts know that the clinician doesn't know. I don't think it's always appreciated that it's only opinion. I prefer to say [when producing a practice recommendation] that there is no evidence of the kind we believe in that allows us to say you should do or not do this practice,” she said in an interview. “There is a tendency to make guidelines and recommendations seem authoritative. I believe that physicians think that there is a great deal more behind authoritative recommendations than there might be when you lift the lid of the box and see what's underneath.”
Dr. Petitti prefers the USPSTF approach, which distinguishes opinion-based recommendations from evidence-based ones, and also labels opinion statements to avoid categorizing them as recommendations.
She also noted that when study results are lacking, the USPSTF often goes through a “chain of evidence” process, an attempt to build a “plausible pathway” from the existing evidence to a recommendation. She gave the example of building an evidence chain to say that weight loss prevents cardiovascular disease. In the absence of direct evidence, the USPSTF would focus instead on documenting the evidence that weight loss improves hypertension and serum lipids, and has other proven benefits that in aggregate establish the broader premise.
Dr. Petitti also finds fault with two other aspects of ACC/AHA guidelines: the involvement of experts with conflicts of interest—a problem that the ACC and AHA attempt to resolve by full disclosure of potential conflicts, and the huge number of recommendations generated.
“I think people are naive about their ability to make unbiased judgments in the face of personal financial or intellectual interests,” said Dr. Petitti, who said she prefers to completely bar people with a conflicting interest in a recommendation from voting.