Thursday, December 21, 2017

The fallacy of expert-centered guidelines

The Health News Review blog posted a story yesterday asking why there was such extensive news coverage of the American College of Cardiology / American Heart Association hypertension guideline but so far, so little of the American Academy of Family Physicians' announcement that it was not endorsing the guideline (the only other news outlets to have reported this decision are the AAFP's own news publication and Medscape). It's a good question, since "de-intensification" (stepping back on intensity or frequency) of recommendations in clinical practice guidelines is relatively uncommon. In this case, the AAFP decided to continue to adhere to blood pressure targets established by the JNC-8 hypertension guideline and its joint guideline with the American College of Physicians for adults aged 60 years or older.

I don't usually read the journal Hypertension, so I only recently came across an editorial (published on June 26) that criticized the ACP-AAFP guideline for the lack of "experts" on its panel. The 30 named authors of this 3-page piece, most of whom I suspect did not meet the International Committee of Medical Journal Editors' authorship criteria, collectively have a list of disclosures of commercial conflicts of interest that takes up nearly a column of text. Those concerns aside, Dr. Franz Messerli and colleagues stated correctly that "guidelines are traditionally scripted by a panel of experts who are intimately familiar with the topic in question." They went on to compare the glittering resumes of the JNC-8 panel ("most of them were indeed true experts") with the more pedestrian ones of the ACP/AAFP panel and asserted that the latter's dearth of expertise did not qualify them for guideline development:

As per PubMed, 3 of the 7 guideline authors have never authored an article on hypertension and 1 has coauthored a single study only. Moreover, not one of the authors of the ACP/AAFP guideline is known to be a hypertension specialist certified by the American Society of Hypertension or of the American Heart Association Council for High Blood Pressure. ... No ACP/AAFP guideline author is currently serving on the editorial board of a journal dealing with hypertension. ... The mere fact that you know how and when to prescribe hydrochlorothiazide does not make you an expert in hypertensive cardiovascular disease.

This argument echoes past complaints by subspecialty groups that the U.S. Preventive Services Task Force (USPSTF), made up entirely of primary care clinicians, could not possibly know what it was doing when it recommended fewer screening mammograms (without any radiologists!) or less prostate-specific antigen testing for prostate cancer (without any urologists!)

The question is not whether subspecialists should be involved in developing guidelines for diagnosis and management of common conditions such as hypertension: of course they should be. The question is, on what level? At the guideline review and revision phase (ACP/AAFP) or in developing the key recommendations? The ACC/AHA selected as one of its guideline co-chairs Dr. Paul Whelton, a renowned expert in hypertension who was the principal investigator for the SPRINT study, the only major trial to show that a blood pressure target of 120/80 yielded more benefits than harms compared to a blood pressure target of 140/90 in a selected group of high-risk patients (as it turns out, even more high-risk than previously reported). In recognition of this intellectual bias, Dr. Whelton was relieved from chairing duties when SPRINT came up, but apparently was able to fully participate in the discussion and vote.

Incidentally, I had the pleasure of meeting Dr. Whelton at a conference 2 years ago and came away enormously impressed by his intelligence and equanimity. But it's very hard for anyone to see something - in this case, harms of lower blood pressure thresholds - if your scientific reputation depends on your overlooking it. Also, in my experience as a staffer for the USPSTF, panel members who did not treat the condition being screened for were able to follow the evidence without being biased by clinical experience. Pediatricians who weren't conditioned to order yearly mammograms or obstetrician-gynecologists who never ordered PSA tests were more often assets than liabilities on these topics.

Finally, primary care clinicians, particularly family physicians, have unrivaled expertise in treating the whole person, not only one body part or organ system. As I wrote in my Medscape commentary on the ACC/AHA hypertension guidelines, "Cardiologists have the luxury of only needing to be concerned with cardiovascular disease, but the vast majority of my patients with hypertension have comorbid chronic conditions and take several medications. Adding one more anti-hypertensive drug means more potential side effects, medication interactions, and costs to the patient and the health system." Kudos to the AAFP for having the courage to stick with the evidence and buck this one-low-blood-pressure-target-fits-all trend.