The marginal value of health care and hospital admission

Here is the job market paper of Nathan Petek, from the Booth School of Business, University of Chicago:

Abstract: The marginal benefit of health care determines the extent to which policies that change health care consumption affect health. I use variation in access to hospitals caused by nearly 1,300 hospital entries and exits to estimate the marginal benefit of inpatient care. I show that hospital entries and exits cause sharp changes in the quantity of inpatient care, but there is no evidence of an effect on average mortality with tight confidence intervals. I find suggestive evidence of an effect on mortality in rural areas and for the over-65 population with magnitudes that imply the marginal benefit of inpatient care is significantly higher for these populations than for the average patient.

Even for rural areas and the elderly, an effect is not seen until more than a year after the event.

By the way, $900 billion is spent annually at U.S. hospitals.

For the pointer I thank David, a loyal MR reader.


Maybe I missed this somewhere in the paper, but isn't it difficult to estimate the effect of hospitalization under Medicare when we're unsure of the other insurance and medical arrangements the patients studied have carried over the course of their lives? Would the benefits of more inpatient care observed in older and rural people be especially pronounced because (respectively) a) they have lived the longest without preventative healthcare b) live in rural areas where their access to preventative care has been limited?

Also rural people are much less likely to go into the hospital than almost any other group because it often involves signifigant seperation and family hardship, so they tend to defer until the last moment.

Seeing grandma and helping her in the hospital is far more dfficult when it is involves a multiple hour drive, and rural people at all income levels are much more accustomed to support from family members and those they have social ties to and much more uncomfortable with impersonal care than urban/suburban/exurban people.

The portion of the US that's rural enough to require a multi-hour drive to a hospital is small. I think there's some validity to your comments, but the US has a lot of hospitals. To be fair though, many of the rural hospitals are going out of business. The trend was already that way, but I believe that Obamacare regulations have expedited the process by cutting reimbursement levels.

Just to point out that there are hospitals and there are hospitals. Small town community hospitals may be a great place to go for general emergency care or to have lab work, diagnostic work, &c (or they have been in my experience), but they're not the place for complex or specialized work. To take one example, cancer treatments in our house required commutes, though not multi-hour commutes (and the practice in question had satellite clinics). Now, university hospital complexes are to be found as a matter of course in any metropolitan region which has a dense settlement of 550,000 - 650,000 at its core. The thing is, you get out into Montana, and the nearest loci that size is Salt Lake City. You in Alaska, the nearest place is Seattle. We have a proximate relation who was taken from a regional hospital via medivac to Sacramento because that was the nearest place she could be treated and they were frightened of sending her by surface transport.

"Just to point out that there are hospitals and there are hospitals."

That's true. There are probably significant areas without specialized care.

And what care you have is really not very good because of it, and everyone living there knows it, and other than the poor acts accordingly.

"By the way, $900 billion is spent annually at U.S. hospitals." Hospitals are the black hole in health care. Recent (10 years) success in bending the cost curve in health care is attributable to the migration of health care services to (much more efficient and lower cost) outpatient facilities. Yet, hospitals are at the center of recent developments in the delivery of health care, as hospitals absorb competing outpatient facilities (and each other through consolidation) and employ more and more physicians (hospitals now employ more than 50% of physicians). A valid criticism of health care reform is that it put hospitals at the center, a choice based not on efficiency but ability to absorb the cost of converting to electronic medical records (among other improvements); indeed, the single most repeated criticism of health care during the deliberations over health care reform was "fragmentation" in the industry. Cynics believe it's part of a plan for government regulation of health care: It's easier to regulate an industry the greater the concentration in it. In my low country community, there is a single (not for profit) health care system that operates the only hospital and outpatient surgery center and employs almost all of the physicians (the exception being providers who are reimbursed mostly outside third party reimbursement). Appreciating the risks of geographic concentration, the health care system will soon become art of a much larger (and geographically dispersed) health care system. And so it goes, not only in the low country but throughout the U.S. [And don't be fooled by "not for profit" status of many of these large health care systems - the better question is not for whose profit.] Is it all part of a plan to regulate health care? Probably not: too often we attribute to intelligence what is more likely lack thereof. Change will eventually come, either through reform or the black hole.

Recent success at bending the cost curve was due to the recession more than anything.

I spent enough time in Asia to realize the overwhelming factor in how much things cost has nothing to do with production cost and everything to do with how much money the seller thinks you have in your pocket. We are no different. This leaves us with the simple realization that the problem with rising health care costs (as with rising college costs) is the fact that we keep coming up with creative methods to absorb the astronomical fees service providers are charging, in turn justifying higher and higher costs. I don't advocate the outlawing of medical insurance/third party payments, but for sure if we did, overnight service providers would be forced to start charging reasonable prices; no more $100 pill of advil, no more $5000 MRIs.

This points to the strong need for states to more stringently regulate entry of of new hospitals through certificate of need-like programs.

Also, more care is of the ambulatory variety these days. Would be interesting to see the impact of those outpatient clinics.

CON programs raise barriers to entry and lead to increased concentration in hospital markets. Increased concentration is empirically associated with higher prices in hospital markets. (I think Dranove has a good paper on this; my recollection of the literature is that there is strong evidence that increased concentration causes higher prices.) Absent price regulation, CON programs are not likely to decrease spending on hospital care.

BTW, the more hospitals = more care (and more expense) is also true for other health care providers, like physicians. There is strong evidence that areas with more physicians (per capita) of a specific specialty lead to greater consumption of that type of specialty care (say, cardiologists and cardiac interventions). Of course, the causality can run both ways, but my recollection of the literature is the evidence is pretty strong that the causation runs from more cardiologists to more interventional cardiology, and not from more patients needing interventions to more cardiologists.

The joke about lawyers is that the only lawyer in a small town had too little work to keep him busy, until another lawyer moved into town and then they had more work than the two of them could handle.

Did they control for price? I would expect that more doctors would equal (at least a little) more competition and reduction of prices, so more of the population would try and access the services. Or maybe it doesn't work that way with healthcare.

Increased concentration in hospital markets happens because systems buy each other up--not due to CON. MA is one state trying to address this:

CON laws increase the returns to mergers and acquisitions by restricting entry of new competitors. It's makes more sense to buy up the competition and consolidate the market if you know that new competitors are unlikely to emerge.

CON regulation is for lazy incumbents. Supply does not create demand, it only lowers price, at the margin. Having been in an area where CON regulation was eliminated, what you see is expansion of clinics doing work hospitals once did, and greater specialization in hospitals to do what really needs to be done on a 24 hour basis. I am lazy at the moment, but there is a deep econ literature on the effect of CON regulation and its relationship to higher costs. The CEO of one large HMO I represented argued that CON regulations limited his ability to threaten to transfer patients from one hospital to another by helping another hospital to expand its clinical services in a specialty in exchange for an increased flow of patients. However, once the HMO acquired a hospital, he argued that CON legislation was necessary to keep the failing hospital he acquired in business, with the beds full.

Jan, I agree that the major proximate driver of increased concentration is mergers and acquisitions. But as others point out, sustained concentration is aided by CON laws. In my experience, CON challenges to new facilities are usually funded by incumbents seeking to block entry of new competitors.

I disagree with Bill that supply does not create its own demand in hospital and physician services. Health care is a quirky market, where (because of insurance), consumers (they people who decide to get care or not) do not face significant costs at the margin. Thus greater availability of providers (greater "access") can lead to an increase in quantity demanded. The total cost to a patient is the out of pocket cost (deductible and co-pay) plus time/travel/hassle costs. Often these latter costs dwarf the former, and an increase in the number of providers lowers these costs.

Since most medical procedures usually are invasive, and may involve some pain, I am always puzzled by those who argue that if you have more supply of doctors you will have more patients seeking incessant probing, pinching, and stabbing and injecting.

"This points to the strong need for states to more stringently regulate entry of of new hospitals through certificate of need-like programs."

Which will inevitably lead to fewer hospitals and lower access to hospitalization level care. So why do we need to do this?

Per the paper, over-supply (beyond what is necessary) leads to increased costs at the margin due to more that does not improve outcomes.

" the margin due to more that does not improve outcomes. "

The paper measures improved outcomes by mortality rates. There's a big difference between saying "the healthcare was cheaper and the death rate was about the same" versus the "healthcare was better".

Or perhaps we should ask why hospital care is one of the few markets where more supply tends to produce higher prices?

Why do you keep plugging American economists, Mr Cowen? As a keen advocate of immigration shouldn't you be telling up about students graduating from foreign universities?

The overall objective is to promote the "Brand" (profession economists, preferably U.S.). Specifics of who they are and what they might actually produce are lesser details.

When Nathan Petek solemnly states: " I find suggestive evidence of an effect...", he is saying absolutely nothing of value. But his empty words must be highlighted as profound to incrementally market the brand.

He hasn't been.

Increasingly, the real question is why plug economists at all?

Clearly we should "nudge" more people to move to rural areas, therefore increasing the benefit they receive from medical care. QED

Makes sense.

Related trivia, I know a couple doctors who won't retire even a few hours away, because they believe you have to be near UCLA for adequate emergency cardiac care.

Standard note that HMOs encourage phone calls over visits, and to the extent that we fund other people's care, we should prefer that model.

My insurer has a nurse line to call for a lot of the little stuff my wife would normally go in for a visit for.

Doesn't this say the market is supplying roughly the right number of hospitals? Maybe a few to many once marginal costs are taken into account (insurance does tend to cause us to push past the point MB=MC). But if we found opening new hospitals saved tons of lives, that would say the market is under supplying hospitals.

"Look, it doesn't matter how much we restrict access, or how long these people have to sit in triage, because no matter what we do, patient mortality doesn't change much!"

When the quality of health care is assessed with things like total-number-of-patients-dead then you reach all kinds of insane conclusions. To me, this says more about the marginal value of health economics than the marginal value of health care.

It does seem rather coarse grained. They don't find anything because they're not actually looking for much of anything.

Yes, these are good points. Measuring the " marginal benefit of inpatient care." by looking at changes in "mortality rates" is using the wrong metric.

Bravo. I wish we'd emphasize this point more.

People living in rural areas have lower life expectancies and it is getting worse

I hope he controlled for the known fact that patients going to urban hospitals are generally in worse shape than those going to rural hospitals. Rural patients travel to urban centers when their health is in serious danger. So the state of patients entering urban hospitals is not comparable to state of patients entering rural hospitals, so outcomes (esp measured purely in terms of mortality) are unlikely to be a meaningful factor of comparison. In other words, rural hospital get the easy cases, and urban hospitals the more complex ones.

Paging Robin Hanson

Comments for this post are closed