How Clinical Guidelines Can Fail Both Doctors and Patients

Any confusion over the recent news of cholesterol guidelines in the U.S. is perfectly understandable. On the one hand, the guidelines suggest that nearly half the population should use statins to stave off heart attacks and strokes. On the other, use of the drugs is not with potential side effects and, to many, will offer no substantive benefits. The controversy highlights a problem mired in an outdated way of thinking about health care and the doctor-patient relationship.

Guidelines came about after generations of physicians wanted to bring something more than “opinion and experience” to the patient’s bedside. In the late 1960s legislation for the U.S. Food and Drug Administration was amended to call for a demonstration of efficacy and an assessment of benefits and risk as prerequisite to the licensing of any pharmaceutical. Modern clinical science resulted, first slowly and now with an avalanche of clinical trials, each pouring forth outcome data galore.

The Burden of Clinical Data

Clinicians are expected to stay current with this wealth of information. The modern medical curriculum instructs all budding physicians on how to evaluate the quality and the clinical relevance of all such contributions to the body of clinical science. Because some (or perhaps many) find this exercise overwhelming, there are organizations—many academic and some without any discernible relationships with purveyors that could pose a conflict of interest—that attempt to bundle the information in a fashion that might be relevant to particular physicians or physicians in particular specialties. Some of this bundling is quite systematic, some quite helter-skelter.

Occasionally there is a contribution to the literature that offers an unequivocal advantage for a particular patient group. More often, the bundlers are faced with a heterogeneous literature that often demonstrates little, if any, efficacy. Faced with these circumstances, biostatistics has offered up many a method to impute more value to the literature than is apparent at first blush. The result is that all this bundling adds to an enormous and ever-expanding secondary literature.

What is the clinician to do?

Some clinicians suck it up and even suck it up in discussions with patients about the level of certainty that pertains to clinical options. Many take refuge in another trend that starts about the same time, the promulgation of “clinical practice guidelines.” There are nearly 3,000 such exercises waiting to extricate the clinician who is faced with responding to a patient’s complaint. These Guidelines are the work product of “thought leaders” who have undertaken the same sort of analysis of science and are charged with declaring the implications of the state of the science. This seems a reasonable compromise between the difficulties any clinician might face in dissecting the literature relevant to a particular patient and the difficulty clinical epidemiologists have in separating the wheat from the chaff in any particular literature. The “Guideline” on the topic unburdens all; it is today’s “truth.”

Guidelines as a Disservice

But for us it is not “truth” but an erosion of the patient-physician contract for an empathic trusting relationship and more often than not, a sophism. We will use the guidelines for treating cholesterol recently issued by the American College of Cardiology (ACC) and the American Heart Association (AHA) as an object lesson. This guideline has been met with considerable controversy regarding its validity. We will not engage that controversy. Rather, we will examine the way this sort of guideline facilitates patient-physician dialogues.

There are many guidelines that could provide similar object lessons: the recent British guideline for stents for coronary artery disease was worthy of a treatment similar to that we are applying here. We will argue that guidelines and the principles that underlie their promulgation serve many “stakeholders” quite well, but they do a great disservice to patients and the 21st-century ethic that demands a patient-physician dialogue that promotes informed decision-making.

Progress in ethically driven medical care has ground to a standstill because physicians, insurers, economists, governments, pharmaceutical companies, medical device companies, hospital executives and other stakeholders currently determine the menu of diagnostic and therapeutic options by more mechanisms than just Clinical Practice Guidelines. By virtue of various indemnity schemes, including Medicare, the choice of menu items is largely predetermined for any individual. Much of this occurs in committees and advisory panels that are tucked into the process of each stakeholder. Since these menus are produced by august bodies, they have the effect of defining the standard of care. Some of these guidelines consider how to do things to patients; these are not the object lesson for this essay. Our focus is on the guidelines that consider whether to do things to patients—guidelines like those from the American College of Cardiology and the American Heart Association (ACC/AHA) for treating cholesterol levels in the blood.

At best, these guidelines are recommendations based on scientific studies with results that pertain to the average among us. They do not adequately incorporate the personal differences and preferences of each of us as individuals. Furthermore, while these recommendations are based on clinical science, rarely is the science complete or incontrovertible. Hence, the recommendations are consensus statements reflecting the perspectives of those charged with the production of the guideline. Of the thousands of clinical practice guidelines that have been produced, the majority is based on inadequate science and therefore reflects the conjecture of the “thought leaders” recruited to the task. Many of the thought leaders admit to relationships with industry that are potentially problematic and all have the human foible of preconception.

How the Cholesterol Guideline Falls Short

The recently published ACC/AHA guideline for managing cholesterol levels was written by movers and shakers in the cardiology world who have been charged by the ACC and the AHA to recommend what might be best. The newest response to this charge results in a recommendation that suggests, to our reading, that nearly half of us should be swallowing one statin drug or another. The participants writing this guideline were faced with robust data that demonstrates a correlation between cholesterol level and clinical outcomes over time. But despite the substantive science behind the connection, the correlation is weak. The guideline neglects to emphasize this fact.

For example, the recommendation is based on the 38 percent reduction in adverse clinical outcomes—heart attack, stroke and other cardiovascular events—that result from treating cholesterol levels. But, this number alone offers no meaningful information for a patient faced with the decision to accept treatment; it is a relative,not an absolute value. It could mean that if my risk of a heart attack or death over the next decade is about 60 percent, I can reduce the risk to about 40 percent by swallowing statins every day for the next 10 years. This reduction in risk is substantial. The 38 percent reduction in risk, however, means less to me if my risk is 1 percent, not 60 percent, over the next decade. At that level of risk, taking statins daily reduces my risk a small amount, only about 0.5 percent— if you believe that this tiny effect is measurable. Each of us would value the popping of statins differently in these two circumstances: one circumstance provides a tangible improvement in outcomes at 10 years and the other a questionable one.  An average amount of 7.5 percent, proposed in the ACA/AHA guideline, means nothing to me as an individual. That 7.5 percent risk is an averaged amalgam of many individual risks. Applying one to all is not advanced health care decision-making

The reality of the average benefit is closer to the second example, a small risk.  The best we can hope for in treating 50 people, even 50 people with higher risk profiles, for their entire lives with a statin is to perhaps spare 1 of those 50 from a fatal heart attack. The other 49 will have been wasting their time, effort, co-pay and deductible. Are you willing to swallow this recommendation in the hope that you are not among the 49 out of every 50 who wasted time and money and risked a far greater likelihood of side effects?

Criticizing the reasoning of this guideline panel or their notion of patient education is not our main agenda. The point we want to make is that there should be no clinical “guideline” for reducing the risk from cholesterol at all. Guideline recommendations become standards of care and lead to treatments being applied in all people regardless of their personal appetites for taking risk, either of not swallowing the statin or of side effects if they did. The guideline developers are not the decision-makers; patients are. The only guideline needed in medical care is designed to allow, even require people to choose their medical interventions after being fully informed as to what is known about benefits, costs and risks. No one, “thought leader” or not, can say what is best for an individual. Making a recommendation that most should be treated denies patients their rightful place in the decision-making hierarchy.

Decisions Should Fall to the Patient

The idea of the patient as the sole decision-maker—and not governments, insurers or guideline producers—may seem like a radical shift in the philosophical premise of making medical choices. But it is this sort of radical shift in the goal of medical care that is needed and is feasible thanks to an informative science. The patient is not the partner in medical care; the patient is the leader, the pilot, the decision-maker. Patients don’t need recommendations; they need information about the added benefit of one therapy versus another balanced against the added harm of the therapy. People are smart enough to make the trade-offs; they don’t need to be told what to do. People are smart enough to understand it is as rational to forego statin treatment as it is to adhere to this guideline. In fact, people are smart enough to apply a similar razor to decisions regarding all sorts of items on the therapeutic menu from knee arthroscopy to cardiac stents.

Changing medicine from a “telling what to do” profession to a “telling the patient the information” profession will lead us to a better system of care, better in every way. The profession of medicine should have the goal to serve and inform, not recommend. (A series of essays recently posted on Scientific American discuss this mandate in detail—see below for the links). When uncertainty prevails, no longer should a patient ask, “What would you do, Doctor?” The question should be “What would you do, Doctor, if you were me?” That’s a question that demands a trusting, empathic treatment act between two people committed fully and only to serving the needs of the partner who has chosen to be a patient.

Robert McNutt, MD has been an associate editor at the Journal of the American Medical Association for 12 years and before associate editor at the Journal of General Internal Medicine. He is a professor of Medicine at the University of Wisconsin and Rush University Medical Center.

Nortin M. Hadler, MD is a graduate of Yale College and Harvard Medical School. He joined the faculty of the University of North Carolina in 1973 and has been a professor of medicine and microbioogy/immunology since 1985. His assaults on medicalization and overtreatment appear in many editorials and commentaries and 5 recent monographs: The Last Well Person (MQUP 2004) and UNC Press’ Worried Sick (2008), Stabbed in the Back(2009), Rethinking Aging (2011) and most recently, Citizen Patient (2013).

25 replies »

  1. Added note: that probably means I would end up with a “concierage” doctor rather than an insurance contracted doc or a medicare doc….as long as it remains legal for doctors in the US to practice “off the grid”.

  2. I have already told my primary care physician that I will not provide him with certain types of information unless he agrees to keep private paper notes (so far, these are theoretical concerns as….knock on wood….my life is straightforward and boring)…..but in the event I should choose to smoke pot or have an extramarital affair or experience depression I do not trust ehrs. He said he would….but that when he retires his younger colleagues might not be so accomodating.

  3. Wow, great comment thread!

    Agree with all that getting patients as involved as possible in decision-making, and individualizing care, is essential.

    I do think guidelines can provide a reasonable starting point, as Shirie notes, but it should be the start, not the end point.

    Now I hope the ACOs will listen to you!

  4. I agree with Leslie – Guidelines suck but cant they be used as the jumping-off point for the patient-physician discussion? “Here’s what the XXX association has recommended for your condition Mrs. Jones. Let’s talk about these options and any others you might have read about or heard about.” If doctors are getting paid and insurance companies are paying based on adherence to guidelines, then we have a problem. The guidelines themselves have been written (hopefully) in good faith that they might help.

  5. Gentlemen:

    Thank you for the thoughtful reply. While you and I might disagree about the “business of medicine” being any worse now than in the past (it’s always been there; it’s just that the dominant players change), I couldn’t agree more that moving the culture of medicine to a true partnership between doctor and patient is the way forward for doctors and patients alike. The phrase I use, cribbed from an article on health IT by Blumenthal about a decade ago, is that it will help rebuild the “moral pillars” of the profession.

    Your clear-sightedness about that need is appreciated.


  6. First, Dr Hadler and I thank Dr Millenson for his comments. Debate and clarification are great outcomes of such a process. His points are good ones.

    But, Dr Millenson missed our point, and, hence, it gives us a chance to reiterate.

    Our blog would be considered a poor evidence-based review of the value of guidelines. Our intent was not to debate the value of guidelines but to say that patients will and can decide the value of the use of tests and treatments better than any group of guideline producers. Patients can understand the value of information about any suggested intervention if the information is presented to them and, hence, in our view, any guideline that promotes or suggests a patient’s plan of care is wrong at it’s core. Decision-making is the patients job; not the job of a guideline producer or follower, no matter how well intentioned the guideline.

    We, too, have been on guideline committees and know too well the process of how decisions are made on those committees. Conjectured thinking and biased judgements aside, we also admit that some guidelines can change care; but those will incorporate the patient’s values, not the medical system’s values.

    Last, while we did not intend to debate the value of adherence to guidelines, we can find only association data of precarious value to show that using guidelines makes patients better. It is a hard thing to study. The AHA/ACA guidelines, alone, number over 2700 and 50% are conjecture; again, tough to study the effects of such a broad portfolio. A review of 91 guidelines in primary care could not find evidence of effect. We agree, there are plenty of studies showing an association of adherence and outcome, but association studies are not patient data studies; association studies should raise questions, not support the value of guidelines.

    It is time to move back to patients and away from the business of medicine; it is time good doctors moved away from generalized thinking to individualized thinking. That was our point. For sure, if we could all agree that patients really are the reason for medical care in the first place and that only the patient should decide, we could step beyond the debates of how well our misguided medical care system works and debate how to make it better.

  7. Gentlemen:

    I am disappointed that individuals with as eminent and respected credentials as you both possess are purveying an inaccurate history of how guidelines came about. They did not arise because generations of doctors “wanted” something other than individual judgment; they came about because the Evidence-Based Medicine Movement forced physicians to see that unaided judgment was inadequate. That, plus concerns about the cost of ineffective treatment.

    The AMA fought them every step of the way, refusing to even use the word “guideline” for years, preferring “practice parameters.” Your article, alas, is in that tradition of glorifying individual physician judgment while playing down its pitfalls — pitfalls, again, documented at great length in JAMA and other publications. As for criticizing the ACC, you also fail to note that it was the profession that moved to control guidelines by putting them in the hands of professional societies — including, again, the AMA in a bit of negotiating with Congress.

    Perhaps you also think that banking laws were demanded by savings & loans concerned about their customers?

    The history above is fully footnoted by contemporaneous documents in my peer-reviewed book and has been published, as well, in the peer-reviewed literature. Interestingly, the AMA played a very positive role in trying to get doctors to voluntarily follow “best practice” through traditional means; i.e., sending out information about beta blockers post-MI. Maybe half of docs did what the literature said they should do a decade after the evidence was clear. With the advent of guidelines and publicity of compliance, that number shot way up.

    Are guidelines often too rigid? Yes. Do they need to be more personalized? Yes. BUT: Are fewer patients dying and being harmed because of guidelines than without them, even in their current imperfect state. Absolutely yes, a fact you do not mention. As a result, it’s hard to evaluate whether your recommendation about cholesterol guidelines should be embraced when your view of guidelines in general is so blatantly one-sided.

    Bravo for your recommendations on a different relationship between doctor and patient. And, maybe, bravo for your recommendations on guidelines. But as you know, the best diagnosis is based on an accurate history and physical. Respectfully, the history here simply does not make the grade.

  8. How much has been forgotten or ignored in the natural history and pathogenesis of cholesterol / lipid plaques. Are we just going to be looking at numbers and ignoring individual risk and pathology? Plaque buildup begins in the late teens ans early twenties. At what point and in who a plaque ruptures is ignored and not addressed in the numbers game.

  9. Disappointed…how are doctor’s supposed to properly diagnose and treat according to patient’s individual needs?

  10. Yes, Kim, it does.
    And it always has.
    And generations of physicians have done the best they can to see if calm can prevail so that dying and death do not scar the memories of the living.

    Maybe we’re better off today than ever thanks to living wills and powers of attorney.
    Maybe not.
    Maybe this is reason we need to read Keats with as much investment as we read about apoptosis.

  11. Excellent article. There definitely is a shift in placing more control and decision making in the patient’s hands. Does this cause a problem, however, when the patient is unable to make the decision doesn’t leave a Living Will? When family members are left to making health decisions and can’t reach an agreement, doesn’t that lead to chaos and decreased quality of care?

  12. Hi Nortin,
    I’m thinking less of Larry Weed’s structured problem list, and more that in his book, he made a good argument for individualizing treatment and management, and for physicians being there to help patients individualize the treatment.

    How to properly document these efforts, and how to create periodic summaries of some type…hard to say. But as a doc who sees geriatric patients with long histories, I’m eagerly awaiting breakthroughs that will make it easier for me to get up to speed on the years of prior care. (A really good narrative discharge summary is a wonderful thing, but seems fewer and fewer people are creating them these days.)

    As for who we are: I don’t love the term “providers” but haven’t found a better way to be inclusive; lots of NPs and PAs are helping patients tailor their care, plus others too…

  13. HI, Leslie
    I’m not a lover of Dr. Weed’s structured problem list because I’m a lover of narrative over bullets. But I agree that we must render the discontinuity of clinical interviews over time more seamless. There is no substitute for each physician feeling a need to meet this challenge in the fashion that best integrates the mandates for excellence in care and for continuity. Any short-cut is a compromise that is likely to have a life of its own, as we are all learning from the various “EHRs” we are being saddled with. Sitting before a laptop, filling in the blanks sacrifices more than eye contact. Billing, audits, “meaningful usage” and the other Orwellian neologisms are not the reason we see patients.

    My own solution involves a lot of scribbling in my paper “shadow” charts followed by a great deal of time dictating (and then editing) a permanent discussion of the clinical issues. In America today, that’s no way to “make ends meet” in private practice or support the enormous institutional “overhead.”. We need an America tomorrow that supports creating a record for the primary purpose of communicating issues of clinical concern, references and all. We are not “providers” or “purveyors”, just physicians.

  14. hi Dr. McNutt,
    I assume you’re familiar with Larry Weed’s work? He is very much for individualization of care and has similar worries to you re guidelines.

    He makes a persuasive case for us (patients, providers) needing good record keeping in order to track a person’s response to therapy and find the right path.

    I myself have found it’s often a struggle to keep up with what’s been tried and how a patient is doing on a given problem, so I agree with Weed that having a good strategy to document what has been done is essential. And this, of course, should be open to patients and probably even kept and controlled by them.

    Look forward to your follow up post; this is an important topic!

  15. I am humbled by the responses. Passion lives for the care of patients.

    I believe that the cottage industry of one MD; one patient is stilll the best model. I practiced medicine as a “longitudinalist”, not a cross-sectionalist. I followed patients to make sure diagnoses were correct (after all, the diagnostic process is one of weeding through wrong diagnoses to the correct one) and never felt it was my job to make a decision for a patient.

    So, I, too, am perplexed why we so often follow the “menus” given to us and where the idea came from that “on-average” was the goal of scientific inquiry. The goal is to maximize individual care and not on average care. It seems as if we now give all the same size and style shoes to patients despite the patient’s variable foot size and style preferences.

    One responder makes an especially great point; how do we get back to where we were supposed to be in the first place. I take those comments seriously. So seriously that I will do my best with another blog to comment fully. We can change this.

    But, the first step to any change is believing that it can happen and knowing, first, where we are going with the change. The patient as the only decision maker is a first step; it will force us to develop the science and information of individuals instead of small, non-generalizable groups of patients that produce information that we can’t use very well across our variable patients. We also have to fully understand what variation means; when it is good and when it is bad. The goal of good medical care is highly variable distributions of decisions made by patients from their perspective.

    Thanks to all

  16. Dear Leslie and Leo,
    I’ve been practicing a long time. I have no doubt that most patients will and can leap to the option of making their own decisions if offered adequate information. The exceptions are usually embroiled in issues of disability determination or other settings where people are forced to prove they are ill.
    The limitation is the time required to establish an empathic, trusting therapeutic relationship. “Health care” is compromised by whatever degree that time is discouraged, penalized or delegated. As we all know, American health policy suffers from decades of unintended consequences and now has another agenda (see my last book, Citizen Patient.)
    Scientific American invited me to produce a series of essays that speak to some of this. Two address the act of informing medical decision making one patient at a time:



    Maybe, some day….
    Nortin Hadler

  17. Well said. Things like the NNT and NNH should be the first things to reach the patient instead of deceptive numbers like RRR. The package inserts for pharmaceuticals are loaded with completely useless information (I believe by intention). Why not answer some direct questions for patients?: What is this medicine doing for me? What are my chances, Doc? Physicians have to really dig for this information. There are good reasons why.

  18. “Furthermore, while these recommendations are based on clinical science, rarely is the science complete or incontrovertible. ”

    This is an excellent point that i don’t see very often concerning medicine. Science is a spectrum. There is hard science like physics where experiments are repeated over and over in an effort towards revalidation and purity of theory. There is other science like medicine where just a few small limited studies can produce “the standard of care”. The studies can’t be repeated, because that is “unethical”. Rolling back poor quality evidence takes decades. Perverse financial interests are constantly confounding what should be honest investigation. Despite the lab coats, medicine leans closer to voodoo than it does physics.

  19. Agree with the authors that creation and use of guidelines is fraught with problems. Also agree that often the absolute likely benefit of following a guideline is much smaller than patients — and even physicians — realize, even if we assume the patient is fairly average and a good fit for the guideline.

    And of course, I’m all for patients getting good information and taking the lead in their medical care.

    Still, I wish the authors would’ve written a little bit more about how to make that feasible. Eventually we’re supposed to get to personalized medicine which will analyze the genome and make better treatment recommendations. But in the meantime: where are patients supposed to get good information on NNT, NNH, and alternatives? Where should the docs get this info?

  20. To paraphrase the bandits in “Treasure of Sierra Madre” :
    Guidelines? We don’t need no stinking guidelines!

  21. Excellent piece. My first experience in watching guidelines created was back in the early 1990s, when guidelines were produced regarding the use of Epo in patients with ESRD. Thought leaders, indeed, shepherded to the table by a major manufacturer’s consultants, to produce a “guideline” bought and paid for by the company. But it was oh so objective and above board.

    The problem now is that the consuming public and commentators such as myself simply believe that no one is trustworthy anymore. Not the government, not journal editors, not pharma/biotech, and certainly not the people on these “expert” panels whose conflict of interest disclosures are a joke.

    The breach of trust in what should be a relationship grounded in it is a sacrilege.

  22. “It’s almost like cookbook medicine.”

    Almost? What do you mean “almost”?

  23. Great article guys. I find it annoying that there is all this talk about “patient-centered” care, but yet there seems to be this opinion from policy makers that we can all follow these guidelines to perfect patient care the same for everyone. It’s almost like cookbook medicine. And furthermore, the physician’s quality measures will depend on how he/she follows the quidelines.
    While I agree with you that the patient (with the physician’s guidance) should be the ultimate decision maker, I don’t see that happening with the way our “healthcare system” is going.