Tenet 6 (Continued): Health 3.0 is Antifragile — The Problem of Too Big to Fail
I volunteer some time teaching gastroenterology trainees at my alma mater. The senior ones seeking a job will ask me what group I’m with. I tell them the group of me, myself, and I. They have a look of “does not compute” on their faces.
I doubt any big conglomerate will be looking to consolidate me. But one of the main effects of Obamacare — which models Health 2.0 — has been the consolidation of health care.
In a previous post I discussed how consolidation is occurring among hospitals and medical practices. All in the name of value. When it’s just the game of mercantilism and crony capitalism in the distorted market of health care.
But there’s something more deeply dangerous about this consolidation, something that gets us back to fragility.
Health 2.0 wants to control volatility. It says if we can make health care more homogeneous, we can reduce costs.
But Nassim Taleb argues that when we have small payoffs, at the risk of large mistakes that can’t be diluted when the system consolidates, we have more fragility.
This is THE danger. We are fragilizing our health care system, in the name of efficiency. And we’re doing it fast, before we fully appreciate the consequences of centralization.
There are several ways this is problematic: the danger of conformity, the misjudgment of risk, and the bigger squeeze effect.
The Danger of Conformity
Health 2.0 is pushing us to conform to “meaningful use” measures, with increasing penalties if we don’t. Measures like cholesterol, blood pressure, and blood sugar are targeted. And our system incentivizes us to take additive steps instead of subtractive steps to control these measures (more on this in a future post). In other words, more medications.
At first it may appear that we’re making progress. The measures look better. We might even be improving short-term outcomes.
But what would happen in such a system if the measures by which the entire quality of health care is based were to become less valid? Or the interventions to control these measures put people in harm’s way?
For example, let’s say you were incentivized to give statins to control cholesterol, in order to prevent a first heart attack. We’ve learned that you have to treat about a hundred people with a statin to prevent one person from getting a first heart attack. For five years.
We’ve also learned that for every fifty people we treat with a statin, one person might get diabetes. And for every ten people we treat, one person might get muscle damage.
But the system wants you to hit that core measure, because evidence-based medicine says so.
About a quarter of Americans aged 40 and up already take statins. With the latest change in the cholesterol guidelines from the American College of Cardiology (ACC) and American Heart Association (AHA), almost half of Americans aged 40 to 75 would be advised to take statins.
Some researchers are arguing that it would be “cost-effective” for two-thirds of this age group to take statins. And this, in an editorial in the Journal of the American Medical Association: “There is no longer any question as to whether to offer treatment with statins for patients for primary prevention, and there should now be fewer questions about how to treat and in whom.”
Are you kidding me?
Taleb calls thinkers like this “fragilistas.” A fragilista thinks he knows what he knows. Through pseudo-sophisticated risk modeling and cost-benefit analysis, the fragilista entrenches his opinions (Taleb terms this the “Soviet-Harvard delusion”).
But what about what we don’t know? The fragilista mistakes the unknown for the nonexistent. And his models don’t account for the nonexistent.
Taleb calls this naive rationalism. And naive rationalism creates fragility.
For example, there are studies that suggest increased aggression in women with lower cholesterol levels and statin use. And some people develop cognitive impairment such as memory loss and confusion on statins.
Researcher John Ioannidis has for years questioned how valid most published medical research findings really are. Some of the most highly cited research studies of clinical interventions have been diminished in subsequent studies, never tested again, or even outright refuted.
Health 3.0 isn’t against evidence-based medicine. It deplores the practice of medicine that’s evidence-ignorant.
But it is vehemently against naive rationalism in medicine.
If a centralized system outputs evidence-enslaved and naively rationalistic medicine, in the name of standardization, and that evidence base is incomplete or even wrong, we have a problem. A major one.
Because of centralization, the mistakes aren’t diluted. And non-diluted mistakes are dangerously amplified. These iatrogenic errors can destroy many lives.
If we have a centralized system that promotes statins and incentivizes this promotion across the board, and we find later that statins do more sinister things, that mistake is concentrated and magnified, and that mistake would affect many lives. And what we originally thought was cost-effective is destroyed by the Black Swan.
Think I’m being a Chicken Little?
When I was in training, it was commonplace to give patients with cardiac risk factors beta-blockers around the time of surgery. This was evidence-based, according to strong guidelines from the ACC/AHA and even stronger guidelines from the European Society of Cardiology (ESC). And if we didn’t do it, our attendings chided us.
The trouble is, the guidelines were based on a study whose lead author, Don Poldermans, was later discredited for research misconduct. And the ESC pushing the guidelines was chaired by…Don Poldermans.
Centralized bodies that have conflicts of interests move slowly. It took two years from the time the Poldermans investigation began, before the ESC retracted the beta-blocker guidelines.
British researchers wrote several articles detailing the consequences of these guidelines. In one they estimated that over a five-year period as many 800,000 people in Europe may have died as a result — with potentially over half of that figure occurring during the two-year delay. That article was subsequently pulled by the editor of the European Heart Journal. The story is well documented by Larry Husten in Forbes.
Seems shady, doesn’t it?
My contention isn’t that guidelines are worthless, or that medical evidence doesn’t evolve. It’s that we’re deifying centralized standards of care by people who commit naïve rationalism and have conflicts of interest for which they pay little price. The price is instead borne by the many who are harmed when that centralized standard is wrong.
Taleb recalls the Greek myth of Procrustes, an innkeeper who wanted to fit visitors perfectly in his bed. So he cut the limbs of the too-tall ones and stretched the limbs of the too-short ones.
Health is nonlinear. And Procrustean medicine is fragile. We are in danger of setting up a health care system that will accentuate big mistakes, especially in volatile conditions.
I fear the ramifications of a “too big to fail” health care system, which like the banking system is in bigger danger of imploding.
For Taleb, antifragility calls for decentralization, not centralization. Only a decentralized system can seek large payoffs while keeping mistakes small. And reduce risks to the entire system because the mistakes are diluted and localized.
Taleb applauds the airline industry for setting up a system where mistakes such as plane crashes, tragic though they are, are independent from each other. They are investigated transparently and factored in to reduce the odds of future mistakes. And as a result, the industry has a stellar safety record.
Decentralization is key in Health 3.0. If we could practice in a health care system that allows us to take small risks independently, and learn from them without fear of inappropriate litigation that makes us defensive, we could antifragilize the system dramatically.
The Misjudgment of Risk
The issue of risk brings up the second problem of centralization: the very nature of risk taking.
In Health 2.0, we discourage risk-taking. We don’t want volatility, we want uniformity. Which ossifies medicine.
But we haven’t eliminated the risk.
We actually end up taking greater risks because we have a near-fanatical belief in statistical analysis and predictivity, in a nonlinear system like health. We pretentiously think we’re smarter than we really are, that we can precisely model such a complex system. But we’ve just taken the risk underground. Where it festers, until a Black Swan event undoes the system.
In Health 3.0, we encourage risk taking. Not with theoretical modeling, but with heuristics — informed rules of thumb.
Here’s one heuristic in Health 3.0: We would have guidelines on when to intervene and when not to — not how to intervene. We wouldn’t intervene when the benefit is small and a Black Swan event would be devastating; we would intervene when the benefit is big, and the risks are small and manageable.
And because uniqueness is an asset, not a liability, the risks we take aren’t all the same. Also, the risk taking is transparent, so that mistakes remain small and inform us as we move forward.
In a decentralized system where intelligent risk taking is encouraged, a Health 3.0 practitioner can through trial and error tinker with best practices while being evidence-informed. And be free enough to make mistakes, but keep them small by collaborating with other 3.0 practitioners on best practices. So the mistakes won’t be concentrated and liable to take the whole system down.
This is risk at the level of the system, but what about risk at the individual level?
Physicians in Health 2.0 at first glance have more apparent security. As employees rather than independent practitioners, they may find the package of a good salary, benefits, and set hours quite enticing.
But as Taleb explains, they’re at more risk than they realize. Much more.
For one thing, physicians in Health 2.0 are becoming commodities in a machine. In exchange for apparent security, they are exposing themselves, knowingly or unknowingly, to being canned when the system is squeezed.
In other words, physicians are becoming fragile to shocks to the health care system. As the system becomes more exposed to an unpredictable event, their livelihoods will be threatened. Much more than if they had remained independent practitioners, who can more easily adapt to changing environments.
And the autonomy they once enjoyed has taken a back seat to conformity to production targets, quality measures, and utilization guidelines.
Overburdened by these constraints, physicians have had two responses: checking out, and burning out.
If you commoditize what a physician does — if you promote a system of shiftwork and automation — you risk the physician abdicating responsibility to this robotic system. Taleb notes that the Federal Aviation Administration figured this out with pilots.
For example, it’s not unusual to see hospitalist-employees who care for patients in shifts — patients they don’t see outside of the hospital — showing less accountability for a patient’s care.
And one of the biggest reasons for increasing burnout in physicians today is the loss of their unique voices in a centralized system. Under the continual burdens of documenting adherence to core measures, dictating to billing codes rather than patient care, struggling with intrusive electronic health record systems, and pandering to administrators concerned with maintaining the patient census, physicians are feeling trapped.
A work setting more predictable, yes — but antifragile, no.
The Bigger Squeeze Effect
I mentioned physicians getting the squeeze when a fragile system is under strain. Taleb defines a squeeze as what occurs when an institution has to take an action, regardless of cost.
This reflects the third problem of centralization: squeezes get bigger with size. The bigger the institution, the costlier the squeeze. And this too is nonlinear.
We are building big healthcare conglomerates, in the name of efficiency and economies of scale. It’s doubtful that this reaps cost savings even in the short term. But these big institutions will be the ones squeezed the most in Black Swan conditions. And that costs even more money.
Look at what’s happening to one hospital system that expanded into my hometown of Houston. And no Black Swan has even happened yet.
Centralization is fragile.
The path to the antifragile is to think less like an ivory tower economist and more like an options trader.
We already know this in medicine. In the next post I’ll discuss how.