Statin Wars: Less-is-More versus Unlimited Medicine 

flying cadeuciiIt is the beauty of evidence-based medicine (EBM) that a scientist can at once be a Pope and a Galileo. His transmutation is as effortless as it is discretionary. If you think you’ve met Galileo – a rebel, a free thinker, a rocker of the establishment – the following week he is a Pope, castigating detractors, censoring critics, and celebrating uniformity. He changes by a roll of the dice. His change is decided by a quirk in hypothesis-testing known as statistical significance. If the p value is 0.051 he is Galileo, if the p value is 0.049 he becomes the cardinal. He is one day a raging skeptic and another day a true believer.

The latest fight between orthodoxy and free inquiry is about the benefits and harms of statins for primary prevention. A review, and an editorial, in the Lancet said the benefits of statins are real, the harms are exaggerated, and skepticism of benefits of statins should be censored because doubt can harm the public who may not take their statins and thus die prematurely. Stated differently, skepticism kills. The lead author of the review once asked the BMJ to retract a study which he felt overplayed the harms and denied the benefits of statins. The editor compared the fear about statins to fear about vaccination. Statin skeptics, like vaccine deniers, are now medicine’s truthers.

It is unclear what will send the skeptics to the Gulag. Is it a denial that statins have any benefits or a quibble about the precise benefits? Will you be sent to the Gulags if you incorrectly say the number needed to treat (NNT) is 150, when it is in fact 100; if you get the confidence interval wrong; or if you under estimate the significance level? The Spanish Inquisitors made it simple – deny Jesus is the son of God and you’re in the torture chamber. The new truths are probabilistic – God is statistically significant with p value perched precariously at 0.05.

The BMJ, at the crossfire of this inquisition, responded to Lancet’s accusation in kind.  The fight between Britain’s two major medical journals is like the recent spat between the leave and remain camps in Brexit – that there was a winner shouldn’t detract from the pettiness, or the close margin, of the dispute. I shall spare you the granular details of the statin war (excellent summary at Cardiobrief and a raucous take by Michel Accad are worth reading). Granularity in EBM is not the same as looking at the night sky through a telescope. You won’t be awed by magnifying the details. However, if you suffer insomnia, the 30-page review of statins in Lancet might be what the doctor ordered.

What are the harms of statins? Aside from remembering to take the pill, which should no longer be burdensome, but a fact of life like wearing a seatbelt or filing taxes, the harms include muscle pain. Muscle pain? Are you thinking what I’m thinking? Muscle pain! BFD! Is modern man, who by now was supposed to have colonized Mars and traveled to Proxima Centauri in search of a new planet, arguing about muscle pain? What an anti-climax!

The dispute doesn’t have the same metaphysical significance as Earth versus the sun at the center of the universe. Galileo’s captors saw in Galileo’s musings the contradictions of their religion. But the contradiction of statins is more divisive than heliocentricity. But let us not falsely elevate the fatigued quadriceps muscle to martyrdom. The statin war is not about skeletal muscle. It is about ossified ideologies.  The statin war asks medicine’s most primal question – what is the role of medicine in society?

It is a war between two movements in medicine, which itself has become a religion with the physician as its high priest. Medicine is divided into two sects. One is the less-is-more movement. The other movement, the antithesis of less-is-more, doesn’t have a name but I will call it the “unlimited medicine” movement. One emphasizes medicine’s limitations, the other medicine’s possibilities. One is short-sighted, the other can stare only at the horizon. One is too willing to press the brake, the other ever eager to press the accelerator. One seeks redemption in data, the other seeks salvation in venture capitalists. One wants to be guarded by budgets and opportunity costs, the other wants to break chronological budgets. One sees man as mortal, the other dreams of transient, incremental, immortality.

The two movements are profoundly similar. The men who fought for Richard the 1st and Saladin in Jerusalem in the Crusades, though fought against each other, all thought they had God on their side. Both the less-is-more and “unlimited medicine” movements believe they have history on their side. Both movements are sincere, overly sincere, yet disingenuous; both can be dogmatic, both are sentimental, both are self-righteous. Both are right, neither is wrong, yet both are wrong.

The less-is-more movement is at least forthright in its objectives. It believes medicine is doing too much. It believes medicine is overtesting, overtreating, overdiagnosing, over stepping, and over intruding in the lives of private citizens. It believes medicine has lost its moral compass. Just like a nosey mother-in-law extracts too much for the occasional baby-sitting, medicine extracts too much for extending longevity by too little.

On the face, the less-is-more movement is the more ideological and self-righteous of the two. This movement has made two strategic errors. It has erred by ascribing to malice what can be ascribed to chance. It believes, implicitly, that doctors do too much because of greed supported by the incentive structure. This naïve reductionism, which can easily be countered, has alienated many doctors. This is a double blow for the movement which has lost not only its key message but supporters of that message. The movement should, instead, have emphasized that medicine is an art, an imperfect art; that the art is minimalism, and minimalism needs judgment and skill; that the art is not uniform and can vary between its practitioners, but that’s ok; that the art can be ruined by diagnostic and therapeutic incontinence. It has not done so because it does not believe medicine is an art but a precise scientific enterprise. It is this belief which is the root of its second, more fatal error.

The movement believes its core principles can be justified by science, that minimalism is scientific. EBM is the oddest of oddest sciences. It is not a science of precision but a science which specifies trade-offs. You win some and you lose some, and EBM, when done right, tells you how much you win and how much you lose. EBM quantifies trade-offs. EBM doesn’t abolish trade-offs. The less-is-more movement doesn’t acknowledge trade-offs. It fails to acknowledge that therapeutic incontinence saves a few, very few to be precise, but few nonetheless. It denies that overuse of CT for pulmonary embolism, for example, saves a few lives from fatal pulmonary embolism who would have escaped the net if medicine were practiced as a fine art. It cannot get itself to say “we overtest and overtreat and yes we help a few but it’s not worth it.”

In dealing with the small, but real, benefits which EBM keeps churning, such as statins for primary prevention, the less-is-more movement stays away from costs but overplays harms, obfuscates benefits with statistical purity, demands data sharing, confounds by reminding people of financial conflicts of interests. I have never succeeded in getting a less-is-more proponent to answer a simple question: what is the acceptable miss rate of fatal acute coronary syndrome in a patient presenting to the emergency department with atypical chest pain? 1/100? 1/1000? 1/10,000?

If the less-is-more movement refuses to explicitly state what is medicine’s floor, the “unlimited medicine” movement denies that there is a ceiling. This movement, too, denies trade-offs but the denial is subtly guarded in the doctor-patient relationship. How far will you go to save one life, I have asked its followers. What should be the number-needed to treat to save one person from a fatal myocardial infarction? 1/100? 1/1000? 1/10,000? Is there a limit? How much should be spent extending life by a month? The “unlimited medicine” movement can’t get itself to say “yes I know this cancer drug means less funding for public education for kids from poor background, but it’s worth the extra 6 weeks.” This is partly because opportunity costs in the US are not explicit. Rather, they’re insidious like closing of safety net hospitals or libraries for children in poor areas, and the increasing national debt.

The “unlimited medicine” movement is the more sentimental of the two. Question them, say they’re doing too much, and they become defensive, and bring out their violin playing the most annoying self-pitying tune. How dare I question the sacrosanct doctor-patient relationship. Only they and their patient can answer whether it’s worth treating pulmonary embolism in someone riddled with metastases – a sentiment I wouldn’t disagree with if the marginal costs of such pursuits weren’t diffused to society.

The dispute between the two movements, which I will call Statin Wars, even though it’s beyond statins, is fundamentally statistical. The less-is-more movement wants fewer Type 1 errors – fewer therapies that don’t work to be adopted falsely. The “unlimited medicine” movement wants fewer Type 2 errors – fewer therapies that actually work to be falsely canned. Statistics have been recruited to fight ideology. There is an ideological divide between the movements. Is the role of medicine to heal the sick or to stop people from being sick? If it is to heal the sick, how far should the healing go? Neither question is easy to answer.

I believe that the role of medicine is to heal the sick, not hound the well. I’m aware that adhering to my tenet literally like a Wahhabi and ignoring it completely like an infidel has problems. My tenet can easily be challenged. If medicine should be confined to healing the sick, am I opposed to screening neonates for phenylketonuria (PKU)? I’m not, which tells you that I can’t stand up for my tenet even transiently. Yet, I can’t countenance that the role of medicine is to screen for lung cancer in octogenarians. Is there no difference in kind, or degree, between screening for PKU in neonates and lung cancer in the elderly?

The “unlimited medicine” movement wants autobahns. The less-is-more movement wants idyllic country roads with curves, speed barriers, and stop signs. You can’t have both at the same place. You can’t vote for both ideologies. You must choose between medicine’s limitations and medicine’s possibilities.

Both Britain and the US are in the midst of Statin Wars. In Britain, “unlimited medicine” will break the NHS. The NHS was not designed to support the hypervigilance required of physicians if they want to eke every last drop of benefit in treating hypertensives. The NHS was barely designed to treat the sick. It does not have the infrastructure to treat “in anticipation of future sickness.” In the US, the less-is-more ideology will gut healthcare as we know it. Healthcare is a rising sector of the economy. Jobs, not just life, depend on healthcare. Give healthcare a haircut and towns will face the same consequences that they once faced when manufacturing departed. The US needs “unlimited medicine” as much as Britain needs less-is-more.

Categories: Uncategorized

8 replies »

  1. You have to know the shape of the null distribution before the p-test can be used. And not everything has a Poisson distribution. Just think of data that never includes zero or has long tails….there are all kinds of data distributions.

  2. #pointMissed. Embarrassing to give untargeted treatment to everyone to help the few who truly benefit. Mastectomy for all before puberty would prevent breast cancer. Ridiculous. Identify who needs what. This would be hilarious (suggesting everyone needs a costly treatment that only benefits a subset) if it weren’t actually happening. The analogy to vaccination does not track. There is no herd immunity to heart disease.

  3. Excellent presentation of how much of medicine is a belief system rather than a science. Belief systems should not be bludgeoned into people with the force of law or economics. They should also not be lightly dismissed. They should however be funded by those who believe, and not society in general. If you want to go to church, go to church. But please, fill your own collection plate.

  4. “if the marginal costs of such pursuits weren’t diffused to society.”

    Though you frame the debate as having only two sides there is a third side and that is asking yourself why society must be so involved to the point of .049 vs .051.

  5. A p-value is simply a (weakly inferential) point estimator like any other, one residing also (typically mid-stream somewhere) within a distribution — http://www.zgznews.com/ a distribution ranging from the “effectively”-Gaussian to the lumpy (or otherwise misshapen) Chebyshev-ist.

  6. No serious practitioner of commercial applied statistics gives a flip about “p-values” (a point apparently, regrettably lost on academic medicine). Stress-tested expected value computation (mostly trended) is where the action is.

    A p-value is simply a (weakly inferential) point estimator like any other, one residing also (typically mid-stream somewhere) within a distribution — a distribution ranging from the “effectively”-Gaussian to the lumpy (or otherwise misshapen) Chebyshev-ist.

    But, hey, the journals will continue to exuberantly crank out p-values amid the accompanying jargon — usually extended to 4 decimal places, no less (like job applicants’ resume GPAs any more).