POLICY/TECH: Harvard business school prof #3 doesn’t understand economics

And following up on Reggie Herzlinger and Michael Porter in my continued series beating up on Harvard Business School professors who don’t understand health care, I move onto a relative newcomer.

CHCF’s President Mark Smith interviews  Mr Disruptive Innovator himself Clayton Christensen in Health Affairs. Mark asks his a bunch of sensible questions about how disruption can work in health care (as compared to computers and high tech) with our different payment structure, concentration of diseases/costs amongst a few people and (implicitly) the long term non-episodic nature of the care required for chronic disease. And what does Christensen do? He ignores the parts that don’t make sense to him. The best example is this extract.

HSAs And The Individual MarketSmith: <snip> After all, if you’re choosing between a $150 doctor visit and a $39 MinuteClinic visit, this choice might drive your behavior and help develop a market for MinuteClinic. But the 20 percent of people who are responsible for 80 percent of the costs are often choosing between a $47,000 hospital bill and a $45,000 hospital bill–neither of which is something they can pay out of pocket. Can you think of anything that will result in similar pressure being applied to the high-cost segments of the markets, where we pay though an insurance mechanism and are likely to continue doing so?

Yes. You have a little bit of a chicken-and-egg problem there, Mark.
Another analogy: When RCA made vacuum tube-based televisions and
radios, most people couldn’t afford them, and they were sold through a
distribution channel: appliance stores. The appliance stores made most
of their money not by selling the TVs and the radios themselves, but
through the repair of the burned-out vacuum tubes in the appliances
they already had sold. Sony came along with this simple transistor
radio. It was disruptive relative to RCA. Sony tried to get
distribution through the appliance stores because that’s where radios
and TVs were sold. The appliance stores wouldn’t give Sony products the
time of day because they didn’t have vacuum tubes in them, and
therefore the stores couldn’t make money if they sold the Sony
products. Thank goodness for Sony, K-Mart was arising just at the time
Sony was trying to disrupt the industry, and K-Mart wasn’t able to sell
vacuum tube-based products because it couldn’t service them. And so a
whole new system emerged, so that it wasn’t just Sony that disrupted
RCA, it was K-Mart that disrupted the appliance stores. Almost always
that’s the way it works. It’s not just a product-for-product or
service-for-service disruption. It’s a system disrupting a system. So
you can see why health policy people get indigestion when all you do is
to say, Let’s substitute health savings accounts for the current
reimbursement system. You just give people money and say, Now, you go
pursue your own health care. And the person to whom you just gave money
has strep throat, and they’re looking at going to a regular doctor’s
office and facing a three-hour wait and then it’s going to cost them
$150, or maybe that person is going to say, I just won’t go to the
doctor. And maybe that’s all right, maybe it’s not, you know? But you
can see why the health policy people would worry about this.
along with HSAs there were companies like MinuteClinic, then you could
see how, oh my gosh, I’ve got my own money and I can choose whether I
go to a doctor’s office and get soaked $150 to get diagnosed for my
strep throat, or I could go to a MinuteClinic and for $39 in fifteen
minutes, they’re going to not only give me a diagnosis, but a
prescription for the medication. You can see how people would actually
be delighted to have the HSA solution and a MinuteClinic, because they
can get the job done cheaper, faster, and more conveniently. So if
there were a totally different system out here that could disrupt the
existing system and enable these types of choices, then HSAs might be
seen as good news. But when we just present it as HSAs in the old
system, it really is a pretty ambiguous thing.


Where was the answer about the people with the huge hospital
bills—the ones who are not paying out of pocket—who actually cost all
the money? Well it doesn’t fit the mold, so Christensen ignores it.
Obviously what we need is a disruptive change in the financing system
for the expensive cases that can incent the cost-efficient innovations
like the transistor radio for those expensive cases. No one can believe
that no matter how well Minute Clinic does (and by the way it’s NOT
getting most of its business from HSA-holding patients as far as I can
tell), it has anything to do about those expensive cases. Of course
being a responsible outfit it refers them straight off to the

Christensen also talks alot about the Toyota kaizen quality
improvement model for hospitals. But that’s been common knowledge in
heath care since c.1990. The reason is hasn’t gotten implemented is
that, unlike Toyota, hospitals don’t make more money by delivering a
better quality product at a lower cost. In fact it’s been well shown
that those that go down that path (e.g. InterMountain, Virginia Mason) find it very hard to get their payers to reward them for doing it. As the WSJ said

With each MRI that Aetna and the employers avoided
at around $850, Virginia Mason lost about $450 in profit. The payment
system of government-sponsored Medicare, which private health plans
also use as a template, tends to reward the big capital expenses of
buying high-tech machines such as MRIs. The more the machines are used,
the bigger profit margin they pack

What does Christensen say about that issue?

So those kinds of innovations would be implemented
in the tertiary care hospitals that care for medically complex
individuals. I think that there are better ways to run those hospitals,
and we know how to do it. We just haven’t implemented these solutions.

No shit,  Sherlock. The reason “we just haven’t implemented those solutions
is that hospitals’ incentives are like those of defense contractors.
They operate in a cost-plus add-on environment where doing something
wrong and re-doing it gets them more money. The model isn’t Toyota,
it’s Halliburton. And like Halliburton it’s very profitable to push the
envelope on doing more and more, and just hope you don’t get caught.

We know that rational business actors in any field will develop and
use the technology that maximizes their revenue according to the
incentive structure in place. That’s where the incentives for
technology innovation comes from, including the disruptive types like
transistor radios. Sony figured out that if they could get a cheap
radio to the market, Americans had the financial appetite to buy enough
that they’d make a huge amount of money off them. There was no need for
a change in business incentives, and transitor radios weren’t
complicated by the fact some of them cost (or some of their purchasers
had to pay) 20 times what others did for the same product. And even if
they had, it would have been OK for the consumers who had to pay more
just not to have one instead. Sony still made bank.

But that is the situation in health care. And the reason
that we have third party payment is because it’s impractical to get the
people who use the most resources to pay for it all directly, and we
don’t appear (usually) to be prepared to let them all just die on the

So the innovation required is in changing that incentive system.

What health care needs is a change to its financing systems to get
it out of the mode where most of the money is paid over like defense
contracting dollars (i.e. no politicians dare investigate or question
its value less they fear the wrath of the contractors), to a system
where someone—consumers, the government, employers—is forced to make a
real choice between a marginal dollar spent on health care and one
spent elsewhere. Enthoven has a system to do that (consumers would be
forced to pay real dollars to buy a more expensive health plan) which
is basically how it works in Holland. The single payer countries have a
system for that—the health budget must be weighed up as part of an
overall government budget that relates directly to spending on other
services and overall taxes. And although they only have a solution for
the cheap, below the deductible stuff so far, even the free-marketeers
are groping towards some kind of a solution for that. (Or at least some
of them like Arnold Kling have mentioned it).

But suggesting that a technological innovation that doesn’t maximize
revenue for anyone, and doesn’t hit at the heart of the problem is
going to change much of anything is just not helpful. So exactly what
value is Christensen’s analysis adding? Beats me. At the least I don’t
find it very innovative.

Spread the love

Categories: Uncategorized

Tagged as: ,

16 replies »

  1. Disruptive technologies don’t fix things overnight. The discount store has been disrupting the department store since 1960. Disruption is a slow and evolving process, but it does have downward pressure on price and improves quality, and it takes a long time.
    The people with high cost health bills are at the top of the value chain buying the most expensive products. The people going to the minute clinics are at the bottom of the value chain. Minute Clinics, at this point in time as we understand their value, have nothing to do with the value chain at the top. Once Dr’s offices get out of the business of doing things that Minute Clinics can do for much cheaper. Minute Clinics will be left to compete with themselves, and they’ll see a slow in revenue or even a drop. At that point they start to look up the value chain to try and pull something else from traditional doctors and hospitals. Something they can do cheaper and good enough to compete with expensive Drs. But, that’s just one player there could be something we’ve yet to see compete higher up in the value chain reducing prices, but that’s like trying to predict the future.
    The people with high cost health bills aren’t going to benefit from the forces of disruptive technology within the next 10 years or more. Hell probably 40 years. That’s why high deductible health insurance is probably the only reasonable answer to that question for the time being. And he says that in his book.
    Until something like a Minute Clinic can start competing with high cost Drs high dollar care will still exist. I think Christensen’s was answering a different question, or didn’t understand the question.

  2. In a recent stay in the hospital, I noticed that it was
    harder to be truthful to healthcare workers that were
    more attractive or perceivably more attractive than others.
    My point is that it’s hard to tell Dr. Katherin Z. Jones
    or nurse Klume, that you’re a sick worn old bastard and their isn’t much left about me that I feel real good about.
    Do the women that have Dr. George Clooney, coronary analyst B. Pitt, or neurosurgeon Cruise,tell them all of their faults? I think not. .
    Study this. I think it needs to be considered.
    Live On,

  3. Matthew, your enthusiasm is, as always, invigorating.
    As you say, changing incentives for care purveyors would (more positively, will) greatly accelerate uptake of valuable health tech.
    Large employers – the end payers for a goodly share of pricey health care – are lumbering in the direction of addressing the issues you raise here:

    “If we reformed how the health care purchaser market was structured so that people/payers were really incented to buy cost-effective higher quality care, then you’d see the right type of disruptive technology/process popping up all over the place. But currently we pay the system to do more and more with no clear value attached….”

    Among the challenges here, as your boy Jamie Robinson notes, few, whether payers, providers, politicians or policy wonks, deign to tackle the thorny matter of what “value” is, clearly or otherwise – which hasn’t stopped everybody and his doctor from inserting “value” into his transformational call to arms.
    Further, and perhaps more important, societally palatable methods of resource allocation/distribution may defensibly be looked at as ‘technologies’ of a sort, themselves, subject to the same sorts of adoption challenges as operating systems or input devices.
    In the meantime, while the incentives dimension gets hashed out, you might enjoy reading “Zipper: An Exploration in Novelty”. It’s not about healthcare, or about incentives per se, but it is about uptake of transformative technology, a process chockablock with nuance.

  4. Good discussion all. However, Barry/JD, in the end customers were happy to buy the new cheaper steel/airlines. It didn’t require a restructuring of their “demand function”–they got something better, cheaper and bought it. Which meant that those vendors expanded and took over the market.
    There is no mechanism–yet–whereby health care customers, especially those spending money on the minority of expensive people, can switch their spending to the more efficient and better product….because for lots of political and structural reasons the end payers don’t care enough about receiving cheaper, better services to overcome the structural impediments against it.

  5. Regarding the emergence of mini-mills in steel production, their initial advantage was much lower capital intensity as a result of their process of melting scrap steel in electric arc furnaces as opposed to the integrated process of mining iron ore and coal, converting coal into coke, adding limestone and other chemicals, etc. to make steel. The mini-mills were also non-union shops. Even though they paid hourly wages that were comparable to what unionized steelworkers earned, work rules were much more flexible which made for superior productivity. They also did not have the legacy costs of large retiree health and pension obligations. Later, Nucor developed the thin slab casting technology to make higher quality sheet steel which reduced the number of rolling steps needed to make finished steel from the continuous cast strand. That really was a disruptive technology which was developed at considerable cost and risk to Nucor.
    Sometimes the established order is changed by a newcomer that brings lower costs. The airline industry is an example of this. New entrants buy used planes cheaply and hire pilots, mechanics, flight attendants, etc. willing to work for much less than the union rate at the large established carriers. When the upstart’s market share was small, the big carriers just let the new players fill their planes rather than cut prices. Eventually, the low cost carriers won so much market share, the old line carriers had no choice but to learn to compete on price with some of them going through bankruptcy in the process.
    In healthcare, according to the California Healthcare Foundation, hospitals accounted for 38.4% of health costs in 1984, but their share shrank to 30.4% by 2004. Less invasive surgical techniques allowed patients to leave the hospital sooner or even have the procedure done on an outpatient basis. New drugs made some hospitalizations unnecessary and shortened others. The number of inpatient beds has been in secular decline for more than 25 years while some markets (like NYC) still have too many beds.
    Going forward, perhaps outsourcing the radiologist’s function to lower cost countries could happen as a result of computer and telecommunications technology. Minute Clinics could displace some more expensive primary care and (hopefully) ER visits. The power of the Internet could be utilized to provide vastly improved price and quality transparency in areas like prescription drug and hospital costs to force more competition and take cost out of the system. The power of information to make doctors more readily aware of who the most cost-effective providers are could significantly affect referral decisions.
    Counterbalancing these forces, on the other hand, are the new drugs and surgical techniques that keep people with cancer, heart disease, diabetes, etc. alive (and incurring medical costs) for much longer than they would have survived 20 or 30 years ago. As one who has personally benefitted from those advances, I would say that’s a good thing!

  6. There has been a great deal written about technology adoption in healthcare generally, with some attention to disruptive technology, so-called. The HBS has taken quite a beating here lately, but they have a very good case study (9-600-076 for the especially curious) of one such disruptive technology: a device from Aspect Medical Systems that helps an anesthesiologist (or an “anesthesia provider”) determine relative level of consciousness in general anesthesia cases.
    The idea that we identify technologies as “disruptive” only in retrospect has a lot of merit to it, but there are some prospective clues. A couple of things to look for (I think, anyway) that give a clue about whether a technology is disruptive, or is a development of current practice:
    1) Does it move an activity from a more hueristic or judgement/experience basis towards a deterministic basis? Does the new technology essentially “explain” some important set of judgements that are difficult to explain, or are unexplained?
    1a) Does it move an activity from a process basis of “routine” to a more data-driven feedback orientation?
    2) Do practitioners in the area see the new technology as a threat to themselves? (N.B. they won’t tell you directly). Do they hold the new technology to higher standards of proof of efficacy than they ever held their existing modalities to? Do they protest that the new technology simply isn’t needed at all?
    3) Does the new technology require great organizational changes to implement: physical space, very differently-trained personnel, new business processes or models?
    4) Did it come from within the established industry, or from outside it?
    I could probably come up with more…

  7. I might be able to clarify things, drawing again on the talk I heard him give.
    According to Christensen, a common form of disruptive innovation/technology is one that it initially seems impractical, does not produce superior quality goods, and is only marginally profitable. An example he likes to use is mini-mills for steel. These could produce steel more cheaply than traditional mills, but it was of lower quality and only came in one or two forms (rebar, I believe). The mills survived at the bottom of the steel industry, producing a commodity product with low margins. But the new process of production that the mini-mills used (don’t remember the details) had the potential to produce high-quality steel in a variety of forms. These low-margin, bottom-rung players were by necessity lean and constantly focused on improving process and quality. The big mills had no economic incentive to do this on their own. The higher-margin players initially didn’t feel threatened because whenever the mini-mills figured out how to do something new that could outcompete the traditional mills (like make I-beams), the traditional mills exited that market and focused on the remaining cushy product lines that hadn’t been commoditized at such a low price-point (sheet steel).
    But as this process gets repeated, Christensen’s argument goes, the old technology gradually gets pushed out of the market entirely.
    To get back to health care and hospitals, the analogy would be that a thriving hospital system simply does not have an incentive to be a first-mover on a radical P4P payment scheme. This is what Matt is saying, and I don’t think Christensen disagrees. The disruptive innovation is only likely to come from a new start-up or a struggling hospital in a competitive market that needs a way to distinguish itself and gain a competitive advantage. But that competitive advantage will not show itself initially, except in very limited segments of the market. Eventually, though, the business model may push the hospital to improve its systems in ways that hospitals paid the traditional way simply aren’t trying to match….until it’s too late and they’ve been outclassed.
    But as I write it this scenario out, it’s clear that a big problem is the existence of sufficient competition. We all know that health care is overwhelmingly local. Some will travel a long way to go to the Mayo Clinic, but the vast majority go to the nearest place, or the nearest one their doctor has admitting privileges to. Aside from a few urban areas, hospitals don’t compete with each other very much. In rural areas, they hardly do at all.
    So, if this kind of disruptive innovation is to happen in health care, it will have to occur in an area where there actually is competition, and at least some organizations desperate for a new angle that will (eventually) give them a competitive advantage.
    But there is no reason to think the innovation, if it happens and has the potential to revolutionize the industry, will spread at all quickly without government intervention. There simply isn’t enough competition overall to spread it. Moreover, physicians may form a big barrier to change to hospitals trying to go for a robust form of P4P.
    In short, I think Christensen’s idea isn’t as obviously inapplicable to health care as Matt makes it out to be, but neither is it as powerful as Christensen wants it to be. And that is supported by the fact that we haven’t had our major disruptive innovation that improves quality and reduces price simultaneously. If the market conditions are there, why hasn’t it happened in 50 years unlike every other industry?

  8. “if the customer has no incentive or no money with which to buy the new disruptive technology (or process), then they won’t. The technology/process will stay largely unused”
    But is that a disruptive technology? Or just another idea that isn’t going anywhere because it’s impractical?

  9. What you’re missing is the difference between latent demand and lack of demand. There is clearly latent demand for a cheap device that plays the currently available stream of music. Was in the 1950s with the tranny radio, is now with the iPod. In both cases 99.99% of consumers didn’t know what the thing was before it was invented, but as soon as it was packaged and priced right it took off…and it was a quantum leap/disruptive technology compared to what came before it (vacuum tube radios/CD Walkmans).
    The point is that it was disruptive to the then current market leaders because it took away their customers.
    As I hope by now you’ve figured out from my recent posts, if the customer has no incentive or no money with which to buy the new disruptive technology (or process), then they won’t. The technology/process will stay largely unused–even if under other circumstances the technology/process might have been a very good idea.
    So if the purchasers of health care were prepared to switch to those providers who’d adopted the breakthrough new technology/process, then those providers and technologies would emerge and quickly win over the market–as has historically been the case in mobile music players.
    The problem is that there is no market advantage to being the disruptive technology/process leader in health care, and doing something better, and cheaper. The customer (end-payer) doesn’t really care, and hasn’t changed how they buy health care in order to steer their dollars to the disruptive technology/process leader. Instead they’ve more or less happily kept buying the equivalent of the vacuum tube radios from RCA.
    That’s why the Kaizen techniques in hospitals haven’t gotten very far, and why Virginia Mason is going around with its begging bowl trying to get its customers to pay their suppliers in a different way.
    If we reformed how the health care purchaser market was structured so that people/payers were really incented to buy cost-effective higher quality care, then you’d see the right type of disruptive technology/process popping up all over the place. But currently we pay the system to do more and more with no clear value attached. Which is why all the innovations that have come on the market and been successful since for example the early 1990s are technology or process innovations that allow more and more imaging/surgery/drugs for which more and more dollars can be charged.
    It amazes me that this is something that the Enthoven-ites, the single payer crowd and the Cato guys all understand (even though we disagree on how to get there), but that HBS professors seem not to consider important.

  10. “So “logical demand”‘s presence or absence has zero to do with their emergence. They’re – y’know – [i]disruptive[/i].”
    I’d go further to suggest that truly disruptive technologies are often recognized for exactly what they are – and for that reason are stoutly resisted because they threaten the incomes of others who have an interest in status quo.
    Further, demand for a disruptive technology almost by definition can emerge only after it’s introduced. If demand were the driver, we’d all still be neighbors of Fred Flintstone.
    Any way one cuts this, IMO your conclusion is correct – – logical demand has little to do with the emergence of disruptive technologies.

  11. Matt writes: “Frankly I’m a little baffled about what’s a disruptive technology — in the absence of some logical demand for it”
    Matt, I’d aver that disruptive technologies are almost always recognized as such only in retrospect.
    So “logical demand”‘s presence or absence has zero to do with their emergence. They’re – y’know – [i]disruptive[/i].
    I’m with you that Christensen sidesteps the role disruptive technologies might play in addressing difficult, expensive treatments that, in the current scheme of things, service providers have such compelling incentives to render.
    I don’t know his work, but he does not appear to lay claim to any deep knowledge of health care in the Health Affairs interview. So, he may simply be unaware of the existence of potentially disruptive technologies in the realm of difficult/expensive health care treatments.

  12. jd,
    Excellent point. Along the same lines, suppose CMS or private insurers were able to develop much more accurate risk scoring techniques intended to predict how much healthcare an individual might need over the next year and the next five years. If the technique proved accurate to within a couple of percent across a population of say, a couple of thousand or more, there should be less incentive for insurers to try to attract the healthy and avoid the sick if they will receive risk adjustment payments that accurately reflect the medical costs that the population they actually wind up with will incur. Moreover, capitation payments for large group practices would, in theory, be a more viable and sustainable business model. Finally, measuring utilization vs outcomes across patient populations with similar risk scores could make it possible to accurately reward cost-effective care and penalize care that falls short on these metrics. Presumably, it would also be helpful if there were outlier payments for the extraordinarily expensive cases, while those cases should also get intensive case management and care coordination.

  13. I saw Richards speak a few months ago, and he flat-out admitted that he knew very little about health care and simply assumed that the same phenomena that applied elsewhere in business applied here.
    So, it’s possible that he doesn’t get the cost drivers and incentive structure of health care, but that those of us who do can apply the notion of a disruptive innovation in a useful way.
    Here’s an idealized example: a hospital system that is struggling in a competitive market decides to insist on being paid only on a P4P basis, with no fees for medical mistakes. It makes a major investment in HIT so that it can monitor performance on a real-time basis according to many different metrics. Initially, this hospital may continue to struggle financially for some time because of the investment and because it can’t command a premium for it’s services until performance improves. However, because it now faces greater pressure to improve performance than its competitors, it improves quality and efficiency at a greater rate than they do and soon can earn a higher return than competitors and steal business from them. This business model could expand as competitors adopt it in order to survive, or as the system expands through M&A, or as new entrants use the model because it is seen to beat the older FFS/DRG model.
    Is it significant that this isn’t really a disruptive “technology” so much as a disruptive payment system plus technological tools to implement it? I’m not sure Richard’s use of the term “technology” is really appropriate in a service industry like health care. It is worth noting that he coined the term by looking at the manufacturing industry.

  14. OK then Vince. Why won’t he answer the question Smith asks? And why is he puzzled that Kaizen-type operations haven’t happened in hospitals.
    Frankly I’m a little baffled about what’s a disruptive technology — in the absence of some logical demand for it

  15. I agree with you on Porter and Herzlinger — they are offering half baked concepts as proposals for systemic health reform (altho I do respect them for advancing the dialogue).
    I see Christensen in a very different light.
    Christen’s theory of disruptive of innovation is not offered as THE solution for systemic health care reform. The thesis of his Health Affairs interview is pretty straightforward — we need to put technology into 2 buckets. One bucket of technologicy raises costs; the other is potentially disruptive and can lower health care costs and improve quality. There is no pretense of a solution for systemic reform.
    The fact that Minute Clinics don’t solve the problem of the uninsured doesn’t make them a bad idea.

  16. Do you think the profs at the Harvard “B” school ever talk to the profs at the Harvard School of Public Health? I find the research and writing from the School of Public Health more reasonable and sensible than the B school’s.
    Thanks I enjoyed the rant. A good rant is so therapeutic.

Leave a Reply

Your email address will not be published. Required fields are marked *