“It’s in the new guidelines” does not always mean it’s time to change practice.

I recently wrote about the ACC/AHA/AACVPR/ABC/ACPM/ADA/AGS/APhA/ASPC/NLA/PCNA Guideline on the Management of Dyslipidemia. My essay pointed out a couple of very specific objections I have to the guidelines. Here, Dr. Lesser describes a broader problem with the new lipid guidelines and extrapolates it to all clinical practice guidelines. Adam Cifu The 2026 American College of Cardiology/American Heart Association Guideline on the Management of Dyslipidemia was released last month. If you work in medicine, you’ve already heard some version of the same four words: just follow the guidelines. I want to push back on that, not because the guidelines are wrong, but because “follow the guidelines” has become a thought-terminating cliché that short-circuits exactly the kind of clinical reasoning it is meant to support. This is not an argument to ignore or dismiss guidelines. The 2026 dyslipidemia guideline was a substantial effort. Its literature review is broad, its clinical questions are relevant, and much of its guidance will help clinicians. But “follow the guidelines” treats an expert consensus document as if it were a verdict. It isn’t. The process is more transparent than you think — and that’s the problem The development of ACC/AHA guidelines is a remarkably transparent process, governed by a detailed methodology on the process. But transparency about process is not the same as transparency about deliberation. The methodology manual describes voting procedures and peer review, but the guideline does not show the range of opinion within the committee, how disagreements were resolved, or whether any members dissented from specific recommendations. What appears in print is a unified document; the uncertainty and disagreement behind it are largely invisible. Anyone who has practiced medicine and reviewed the literature knows that you can ask 3 doctors for their opinions on treating a patient and get 4 answers. Uncertainty is a part of medicine, but is obscured by a guideline conclusion. Who’s in the room — and why it matters The ACC/AHA guideline process draws heavily from the cardiovascular community. Writing committees are composed largely of cardiologists and lipid specialists, alongside pharmacists, nurses, dietitians, and a small number of generalists and patient representatives. This is a governance choice that has predictable consequences. A clinician managing complex lipid disorders in a subspecialty setting operates in a fundamentally different environment than a primary care physician managing multiple chronic conditions in a 15–20 minute visit. That difference directly shapes what recommendations are feasible. The imbalance is not just about the care setting. Determining whether a trial demonstrates clinically meaningful results is a clinical and methodological judgment. It requires statistical fluency, familiarity with bias, and comfort interrogating uncertainty. Specialists bring clinical experience that includes the consequences of undertreatment. That perspective is valuable, but it can introduce what is sometimes called specialty bias. Specialty bias encourages physicians to preferentially recommend treatments aligned with their own training and practice. For example, surgeons are more likely than nonsurgeons to recommend surgery, and these recommendations influence patient choices. A more balanced committee would include greater representation from clinical epidemiologists, biostatisticians, and evidence-based medicine specialists — individuals whose primary expertise is in evaluating evidence rather than treating a specific disease. Their relative underrepresentation is common across guideline development. The length problem The 2026 dyslipidemia guideline, according to its authors, is a “comprehensive one-stop shop.” That description is accurate — and revealing. The document is accompanied by slide sets, population-specific summaries, pocket guides, and plain-language versions. This is not just accessibility design; it is a recognition that the core document is too large (134 pages) to serve as a primary clinical tool for most practicing physicians. There is no plausible workflow in which a busy primary care physician reads and integrates a document of this scope into routine practice. The practical result is predictable: clinicians either follow simplified versions embedded in clinical decision support tools without understanding the reasoning behind them, or they disengage entirely from the full document. Not all recommendations are equal The ACC/AHA classification system distinguishes between the strength of recommendation and the quality of evidence. Some recommendations are supported by multiple randomized trials (Level A). Others rely on more limited data or expert opinion (Level B or C). Many of the newer guidelines are “B-NR”, meaning there has been no new randomized trial evidence to inform the change in guidance; they are based on a combination of observational data and expert opinion. A Class I-A recommendation and a recommendation based primarily on expert consensus are fundamentally different forms of knowledge. Yet “follow the guidelines” flattens these distinctions into a single instruction. The most important recommendation is the one often ignored There is an irony embedded in the “just follow the guidelines” framing: it overlooks one of the guideline’s central principles. Shared decision-making is not an afterthought in the ACC/AHA framework. It is explicitly emphasized in the methodology and embedded in the guideline’s overall approach. The process of risk calculation, personalization, and recommendation is inherently collaborative. That collaboration matters most in the gray zones — where evidence is uncertain, benefits are modest, or patient preferences vary. A patient with a 10% 10-year ASCVD risk who strongly prefers to avoid daily medication has a legitimate claim on the guideline, just as one who prefers aggressive risk reduction does. In that sense, “just follow the guidelines” is itself a failure to follow the guidelines. The document does not replace clinical judgment or patient preference; it is designed to inform both. However, the guideline itself offers little help in shared decision-making. For instance, it is sparse on Numbers Need to Treat (NNT) and does not mention that for a patient with a 10-year risk of 15%, the NNT is 25. Thus, most patients who take a statin will not benefit. When guidelines matter most — and when they matter least The usefulness of guidelines tracks closely with the strength of the underlying evidence. When evidence is strong, guidelines are less important. Clinicians have access to the data through primary articles, systematic reviews, and continuously updated clinical resources. Where guidelines matter most is where evidence is weakest — when clinical questions are real, but trials are limited or absent. In those settings, structured expert opinion is genuinely valuable. But this is also where transparency matters most. If a recommendation rests primarily on expert consensus, readers should know how that consensus was reached. Formal methods such as the RAND/UCLA Appropriateness Method or Delphi processes make disagreement visible and quantify uncertainty. Many guideline processes do not report using these approaches, and rarely make the range of opinion explicit. An uncomfortable conclusion The ACC/AHA is among the more transparent guideline-producing bodies in medicine. It publishes its methods, discloses conflicts, and makes a genuine effort to standardize its process. And yet, even here, familiar concerns remain: committee composition that leans toward intervention, evidence of varying quality presented under a unified framework, documents whose scale limits real-world usability, and limited visibility into how expert consensus is formed when evidence is insufficient. Primary care physicians are expected to navigate guidelines from many specialties — cardiology, oncology…