Lost in the Signal

I finally got around to reading Bryan Caplan's The Case Against Education: Why the Education System Is a Waste of Time and Money.  In it, Dr. Caplan, an economic professor at George Mason University and self-avowed libertarian, argues that, aside from basic literacy and numeracy, our educational system serves less to educate and more as a way to signal to employers who might make good employees. 

Oh, boy, did this book make me think about our healthcare system. 

Credit: UCLA Health

Dr. Caplan's views on economic signaling are by no means out of the mainstream, although his application of it to education may be.  Think of it this way:

• how many of the courses you took in high school and college have actually proved useful to you in your career?
• Indeed, how much of what you learned in those courses do you still remember?

Dr. Caplan cities a raft of statistics to support his point-of-view, including ones about adults' dismal knowledge of most subjects after they graduate, how few college majors actually train many people for jobs in those fields, and even how the "learning to think/learning to learn" arguments are not well supported by data.

He believes that employers prefer people who get a college degree (or at least a high school diploma) because the degrees signal that graduates are "intelligent, conscientious, and conformist."  That is, if you are willing to endure the boredom of all those years of classrooms to the end, you're not only at least moderately smart but also are more likely to be willing to endure the tedium that their jobs will probably have.

E.g., if you are hiring and have two candidates with similar grades and courses, but one of them didn't finish a class in their senior year (and thus could not graduate) versus one who stuck it out for the diploma, how likely are you to risk hiring the dropout -- even though, in terms of what they supposedly learned in college, they were almost exactly the same?

Dr. Caplan notes from his own experience how many students search for easy courses and easy professors, and often fail to even show up for classes -- not behaviors one would do if trying to maximize their learning.  He argues that it has never been easier to find great education: you can audit classes for free at many universities, and take online classes from some of the world's best teachers.  But neither will be recognized by employers in their hiring. 

Dr. Caplan offers drastic remedies, such as cutting all government funding to education, making college more expensive, cutting most majors (he does like practical majors such as statistics or engineering), and more vocational training.  He also supports child labor, as he thinks it is better that kids at least learn how to work instead of wasting time in school.

I suspect Dr. Caplan staked out an extreme position to make a point, but that does not mean that his point doesn't have validity.  Most people would tend to support ever more spending on education -- free college for everyone! -- but, at some point, there have to be diminishing returns, and it is certainly possible that we have passed that point.  It's not like we're doing better on most measures of knowledge or skills. 

Credit: Lown Institute

We spend over a trillion dollars annually in government (federal, state, and local) support for education, which is more than we spend on defense but way less than we spend on healthcare.  As with education, most of us would probably acknowledge that healthcare is full of waste -- not just excess administration but also unnecessary/inappropriate care. 

Even worse, all that healthcare spending seems to buy us declining mortality increasing morbidity, and an array of quality outcomes that rival some third world countries.  Perhaps over-reliance on signaling is one of the reasons for this. 

For example:

Degrees: The gold standard in healthcare is the M.D. (or D.O.).  We look for it, we rely on it, we have faith in it.  And we expect physicians to spend more time in training than other countries, and spend more money on it.   

But we don't really know which medical schools (or residency programs) are better than others.  We don't know how individual physicians did in those programs.  We don't know the relationship between excellence in training and excellence in practice.   

Even worse, we don't know which physicians are good at what they do, much less who the best are.  Having a license or even board certification not only does not guarantee basic competence, it doesn't necessarily mean a physician isn't impaired, has lost their license elsewhere, or even has sexual assault charges. 

We like the degree signal so much that we're increasingly expecting nurses to have B.S.N.s (or Master's), and Ph.D.s for nurse practitioners, pharmacists, and physical therapists.   

Does any of this necessarily make healthcare professionals better?  How would we know?  Or are they just signals to reassure us?

Reputation: You know the Cleveland Clinic.  You know the Mayo Clinic.  You know the most prestigious medical institution in your region.  But you probably can't say why they are more prestigious, and they probably can't really prove that they should be.  

We base these views a lot on faith, on word of mouth, on anecdotes, on which have newer/fancier buildings, and, increasingly, on advertising.  On signals, in other words. 

Latest: We like the newest drug, even though it usually is (much) more expensive and may only offer slightly, if any, improved efficacy.  We like new tools like gamma knives or robotic surgery because, well, they are newer and must be better, expensive be damned.  We get MRIs even when CT scans or a simple X-ray might suffice.  

Given a choice between older treatments/technology or newer ones, who among us wouldn't prefer the latter, even when the latter has not demonstrated its superiority, especially in cost/benefit?  The fact that it is the latest is our signal that it must be better.

I don't go as far as Professor Caplan and say that most of our healthcare spending is wasted and should be cut, but I do think we spend more than we should, and too much on the wrong things.  We'd get a bigger bang for our buck by investing more in public health and less in direct medical care. 

We've gotten lost in healthcare's signals, and aren't focusing enough on what they're actually supposed to be telling us. 

The End of Health Insurance

Paul Tullis has an interesting article in Bloomberg about how self-driving cars might kill auto insurance "as we know it."  After all, if human error is responsible for 90% of auto accidents, and those humans are taken out of the equation, what's left to insure?

Many people don't think much about autonomous vehicles, but Mr. Tullis reports that Michelle Krause, an Accenture insurance expert, says that their impact on auto insurance "...comes up in every strategic conversation" within insurers.

It made me wonder: what would it take to kill health insurance...as we know it?

I.U. professor Rodney Parker told Mr. Tullis: "Liability is likely to migrate from the individual to the manufacturer and the licensers of the software that drives the AV."  This means that, as David Ross Keith, an MIT professor, also told Mr. Tullis: "It’s foreseeable that insurance is a much less consumer-facing industry in the future."

The experts that Mr. Tullis talked to see closer relationships between auto makers and auto insurers; as for potential mergers between the two, Ms. Krause told him "those conversations are going on as we speak."

It makes healthcare mergers like CVS-Aetna or Cigna-Express Scripts look far-sighted.  After all, UnitedHealth's non-health insurance subsidiary, Optum, has already become a powerhouse for the company. 

CVS prototype Health Hub.  Credit: CVS

Let's think about what health insurance is for:

• Averting losses: Like every other form of insurance, health insurance was originally about protecting people against unexpected, catastrophic losses.  It still serves this purpose, and should you ever have the misfortune, say, to have a premature baby or to go through a long program of cancer treatments, you'll appreciate this aspect.  
• Budgeting: Americans somehow came to expect that their health insurance should protect them even against expenses that were moderate and predictable -- most notably, preventive care.  It's not that most of us could not afford these services, we'd just rather finance them via our health insurance premiums rather than budgeting them ourselves, which is a crazily inefficient way to do so.  
• Subsidization:  ACA instituted a program to subsidize cost-sharing for lower-income people via insurer payments (although the Trump Administration has been trying to end them).   As a result, health insurance has become the de facto mechanism to transfer money for health expenses to lower-income people not on Medicaid/Medicare.

I'll take these in reverse order:

Subsidization 

Health insurance is not the right mechanism to do wealth transfers.  It's not what it is designed for, and it is not what it is good at.  Such transfers are a social problem, and should be dealt with via the tax code and/or social welfare payments (as we do, for example, with payments for food or housing).   

Budgeting

It's a failure on our healthcare system's part, that we feel we need financial incentives to get preventive services.  Either we've failed to convince people that getting such services is in their best interests, or we've been promoting services for which that case is unclear.  Neither of these is good.  

We need to stop expecting health insurance help us budget for expenses that, in any other aspect of our lives, we'd be paying for ourselves.  

Averting losses

Even if we accomplished both of the above, health insurance would still probably not look too different than it does now.  Our healthcare system would still have catastrophic expenses, and we'd be looking for protection against them.  We'd still have networks, negotiated prices, and tensions between those who deliver health care and those who pay for it.

Credit: WSJ

We have to attack the root problem, which is not just the prices, but also the costs.  Some examples of how this can happen:

Virtual care will allow us to get advice and even treatment where/when we want it, and increasing reliance on A.I. rather than human expertise will both cut direct costs and, hopefully, unnecessary treatments.  

DIY health is a trend that has promise to greatly impact costs.  Whether it is hearing aids, insulin pumps, or "biohacking," we're starting to move away from reliance on expensive solutions from traditional healthcare sources to cheaper, even home-grown solutions.  

Robots, right now, fall within the "more technology, more expensive" ethos of our current healthcare system, but that cannot last.  Robots will get smarter and more versatile, we'll get better at building them, and they'll allow us to take costs out of healthcare in the way they've taken costs out of manufacturing. 

Hospitals, are, as I've stated previously, "19th century institutions operating under 20th century business models in the 21st century."  

We need to move to a future that is not institutional.  We need to move to prevention, to addressing root causes of health problems, and to delivering more care at home and in the community.  With better real-time monitoring, we can do this cost-effectively.

Prescription drugs are one of the biggest pain points for consumer healthcare spending.  Part of this looks a lot like greed, such as seemingly exorbitant increases for previously affordable drugs (e.g., insulin), part of it is the U.S. not negotiating prices as other countries do, and part of it reflects the long pipeline for drug discovery and development.  The former two are more price issues than cost ones, but the latter one is one 21st century technology can help address.

Credit: Chemistry World

For example, we are already in the era of 3D printed prescription drugs, and this will rapidly advance, even to the point of printing your own drugs at home.  This will have huge impacts on manufacturing and distribution costs.

We are also early in the era of using artificial intelligence to aid the drug development process.  A.I. can sort through vast amounts of data to identify likely combinations and monitor side effects, among other things.  The FDA is encouraging such uses.  Novartis, for example, sees itself as a data company, according to Business Insider, seeing A.I. as its "next great tool" in drug development.  

Long story short: take the big costs out, as is possible, and the need for health insurance goes away, or greatly lessens.

We shouldn't accept the status quo; not in how care is delivered, not in how much care costs, and certainly not in how it is financed.  If auto insurers are discussing merging with automakers, Apple is thinking about its post-iPhone era, Ikea wants to become the "Amazon of furniture," and Amazon's own future may be more about cloud computing than retail, then certainly health insurers should be looking to a very different future. 

Virtual Docs.for All!

There are lots of big ideas being pushed these days, in partly reflecting the run-up to the 2020 elections.  There is the Green New Deal that aims to move the U.S.to a "secure, sustainable future," including (but by no means limited to) moving away from fossil fuels by 2030.  In healthcare, there's the evergreen "Medicare for All," which didn't secure Bernie Sanders the 2016 Democratic nomination but which is popping back up (although skeptics point out it is more like Medicaid for All). 

We'll see how either of those play out, especially with our divided, hyper-partisan elected officials, but instead I'll propose something that is potentially more achievable: let's ensure that every American -- at least, those with access to a smartphone or computer -- has access to their own virtual physician (both human and A.I.).

Baidu's Melody.  Creit: Baidu

Healthcare reform in the U.S. in recent years has tended to be about improving the number of people with health coverage (even ACA left some 27 million without coverage) or using more "value-based" payment mechanisms, which have yet to really shake-up our fee-for-service system.

We are a long way from getting consensus on covering everyone, or on revamping the methodologies for the incomes of all healthcare professionals/organizations, but virtual care for everyone is something we can achieve now, at a reasonable cost.

Andy Kessler, writing in The Wall Street Journal, says "it's time to fire your doctor!"  He lists many of the ways consumers can use technology to track their own health, as well as seek medical advice through "several smartphone-based platforms now function like Uber for doctors."  (He is quick to note though, that when you're really sick you'll still want a specialist).

Mr. Kessler points out: "Doctors don’t scale, so the real future of medicine is digital diagnosis...The revolution is coming.  But not from your doctor."  

Credit: Accenture

A.I. has been looming for healthcare for a few years now, with uneven results.  Babylon Health, for example, pitched its chatbot in the U.K., but ran into concerns from physicians about the advice it was giving.  In China, Baidu has been using its chatbot (for both consumers and doctors) for a couple years.  China views A.I. as an essential component of its future healthcare system.  

Meanwhile, A.I. continues to get better. A new study (from China, not surprisingly) demonstrated an A.I. that was highly accurate in diagnosing several common diagnoses.  It's accuracy rate challenged the highest accuracy rates of human physicians.  The thing to remember, as Mr. Kessler noted in his article, is that "The best doctor sees one patient at a time, but a clever piece of code can be used by countless people." 

Studies have also found that A.I. can detect heart disease and lung cancer more accurately than human physicians, and in diagnostic imaging.  There have already been FDA approved AI tools for ophthalmology and stroke detection.  

Healthcare Bot use Credit: Microsoft

In a recent Insights, Forbes cites study that indicate private sector investment in healthcare will reach $6.6b by 2021, with AI generating potential U.S. annual savings of $150b by 2026 (Accenture) or $269b (McKinsey).   IBM (Watson) and Google (Deep Mind) have invested heavily in A.I. for healthcare, and Microsoft just rolled out its Healthcare Bot Service to help spur creation of chatbots and A.I. personal assistants.  

As for virtual care from humans, it is not entirely clear that virtual visits do replace in-person ones, nor are most physicians involved in providing them, but the promise remains.  Kaiser Permanente, which has also long been an advocate of using A.I., claims over half of all visits are virtual.  Many, perhaps most, large health plans/employer health plans give their members the choice of virtual visits.   

Rock Health's 2018 Digital Health Consumer Adoption survey found continuing increases in the percentage of consumers using such services: 75% have used at least one telemedicine channel.    

If we think about it, it is crazy that when we have a health problem, the thing we are expected to do is to call our doctor for an appointment, then drive there (or to the ER/urgent care) to find out what is wrong.  This is why we end up sharing our germs with a room full of other sick people, waiting for our turn to be seen, and worrying the whole about what might be wrong and is going to happen to us (not to mention how much it will cost). 

There are too many unnecessary visits, resulting in too many unnecessary tests/procedures.  Everyone agrees on the problem, but few healthcare professionals think they are the problem.  They're just as happy to have those in-person visits, regardless of the time and money they cost their patients.  That is not putting the patient first, and that is not patient-centered care.

In most cases, the first step should be an AI-based triage to determine what is most likely to be wrong, if I need to communicate with a human physician immediately/virtually, or if I need to seek in-person care -- and, if so, from whom/where, taking into account any health plan network restrictions. 

The U.K's National Health Service reportedly pays Babylon Health about $80 per patient annually, which would equate to around $26b.  That sounds like a big number, but it's probably more like a CHIP-sized bargain.  It's much cheaper than Medicare-for-All or the ACA premium subsidies.  The cost could most likely be reduced by economies of scale, tougher competitive bidding, and, most substantially, passing off costs by requiring health plans to cover it. 

Such coverage would also let the U.S. set the standard in "licensing" AI for healthcare and end the crazy-quilt licensing of human telemedicine physicians we have now.  Both would be boons for the U.S. healthcare system. 

We're already (finally) getting more worried about China's efforts in A.I. -- as demonstrated by President Trump's recent A.I. order -- and this initiative would help us keep pace with their A.I. healthcare efforts.

Giving everyone access to a virtual physician is a good thing to do.  It's the right thing to do.  And it's not only something we can afford to do; it's something we can't afford not to do. 

Let's Stop Healthcaresplaining

You've probably heard of "mansplaining," usually used to describe men explaining something to women in a condescending and/overconfident manner, and often on a topic that the woman knows more about than that man.  Well, healthcare has its own version of that, which I'm calling "healthcaresplaining."

 Healthcare (and health) is complicated.  Healthcare can often be confusing, even scary.  Life and death decisions sometimes have to be made in milliseconds, with no time for discussion or debate.   All those are reasons why many -- especially healthcare professionals -- are adamant that patients aren't and never will be "consumers." 

That's an example of healthcaresplaining. 

The 21st century is a complicated time.  We are surrounded by, and dependent on, technology that we don't really understand and that most of us would be at a loss to repair.  That's why we have the Genius Bar and the Geek Squad. 

It's not just the technology; many aspects of our lives are based on rules and processes that we also don't understand.  As a result, we delegate many of those aspects to "experts."  Something is gained, to be sure, but something is lost too.

For example:

• If you are arrested or otherwise get involved with the judicial system, you almost certainly will hire a lawyer (even if you are one yourself);
• If you have or hope to have significant income/assets, you probably have a financial planner and a tax accountant;
• If you buy a house, you usually use a realtor;
• If you have health issues, you are likely to have a physician, and other healthcare professionals.  

The legal, financial services, tax, real estate, and healthcare systems have each evolved to the point that laypeople don't expect -- and aren't expected -- to understand them.  They and other parts of our lives are growing increasingly more complicated, making us ever-more dependent on those experts.

The thing is, this state of affairs is not preordained.  It doesn't have to be this way.  The truth is that this complexity serves the professions that revolve around it.  Lawyers draft the laws, tax professionals design the tax code, and healthcare professionals create the operating structures of our healthcare system.  

A good example of this are the new requirements for hospitals to list "prices."  Instead of promoting transparency, they reveal the convoluted charge structure behind hospital bills.  They use terminology most people don't understand, is at a granular level most won't actually get charged at, and reflects gross charges rather than negotiated prices.  It's like looking at the SKU of every item used in making an automobile and trying to figure out what you'll pay for the car.

Credit: Caitlin Hillyard/KHN illustration; Getty Images

ICD-10, CPT, RBRVS, HCPCS, ICF, NDC, DSM -- these are all examples of systems that are used in healthcare to describe your condition and/or treatment, in an effort to diagnosis and bill.  They are all monstrously complex and growing moreso.  

We can keep making them more complicated.  There are reasons why we do.  That doesn't mean we should.  

Instead, we should be looking for ways we can stop disintermediating people from the things they are trying to do.  We should be helping people be their own experts, rather than relying on more, and more specialized, experts.

For example, it used to be that the "expert" in buying a car was the car salesperson.  You might try to negotiate with him/her, but there was a huge information asymmetry.  The internet has largely leveled that; the car salesperson still has the edge, because he/she does it every day, but it is at a least a fairer negotiation.

Or there are products like TurboTax, which uses software to let you replicate the tax experts.  Input a few numbers and it can produce your returns, even file them.  In financial services you can go online to buy and sell stocks, bonds, and mutual funds, using various tools to help you pick the best ones for your risk profile, all without any intervention of a financial planner.   

Similarly, there are now services that let you skip using a realtor, allowing you to view/list property online and using an attorney to oversee the paperwork -- all at a substantial discount over a realtor's commission.

Then there's healthcare.  Despite the massive amounts of health information available on the internet -- sometimes referred to as Dr. Google -- in receiving healthcare services we are nowhere near even the parity we might feel with a car salesperson in buying a car.  We don't know enough and we understand less.    

Indeed, healthcare is going the other direction, with calls for "social prescribing" or "exercise prescribing" as examples of the healthcare system trying to take its influence further into our daily lives.  It is true that much of our health happens outside the healthcare system, but that doesn't mean they should become part of it.

Instead of being cowed by healthcare professionals' expertise, we should be thinking of this great quote:

Otherwise, it's just healthcaresplaining.

Previously, I wrote that healthcare has to "do simple better."   I still believe that, but it is necessary, not sufficient.  It needs to make more things -- a lot more things -- simple.  

Tom Vanderbilt had a great article in The Atlantic about "reverse innovation" in healthcare.  Reverse innovation is "taking a technology or solution born of the resource constraints in developing countries and adopting it in wealthier ones."  Some call this "popsicle-stick" thinking.  

As he concludes:

But, in a country like the U.S., faced with spiralling health-care costs and access-to-care issues, where innovation typically leads to more expensive and sometimes unnecessary technologies, it may be time for medicine, still often dominated by a closed, guild-like mentality, to think more inventively. Home Depot might not be a bad place to start.

Healthcare could use "reverse innovation" in so many ways, in the broadest sense of the term.  We should be looking to make the healthcare system simpler.  We should be using language that ordinary people understand.  We should make encounters with the healthcare system less scary, and certainly less frustrating.  We should treat us like our health is our business.  We should help make health our business, and us the experts.

This kind of change will come from both purposeful design and the availability of self-service tools like AI, and it is inevitable -- although neither quick nor easy.  

Enough healthcaresplaining: we are not stupid; it is the healthcare system that is stupid.