« Fairness vs. Goodness | Main | Who Are Macro Experts? »

February 22, 2009


You make a good point, Robin_Hanson, but what about the role of psychosomatic effects? Since a patient's optimism and expectations can influence the outcome (e.g. the placebo effect and the effect of perceived hopelessness), couldn't the patients at the high-status hospital be getting better results than if they had to go to hospital of which they had lower expectations? Is it possible to call out the Emperor's new clothes without worsening average outcomes?

I just switched between medical groups in my HMO in L.A. from one where my wife would deliver our first child in a hospital thats site boasts of a new coat of paint and the addition of a feedback/comment box to the maternity ward, to a group in Beverly Hills that most people would recognize by name from stories about Hollywood stars' ailments.

This change was spurred by Internet reviews and our experience at the first group's GP (2 hr wait/sole whites in the waiting room/office receptionist invisible and un-summonable behind frosted glass/office hawking a line of creams, poultice's, etc). In Bev. Hills, the OB office (of an HMO group)is bright, cheery, efficient, and has a high-tech veneer).

I was initially shocked to be relegated to the first, and then after seeing the first, I was surprised I had access to the 2nd (we cannot switch to PPO until after the baby) even with a 30 minute drive.

I don't know what objective statistical difference it would take for me (and more important, the missus)to switch back.

But now I will at least do a little research. (any suggestions where to look, Retired Urologist?)

Quote from the final page of the article that is the basis for this post:

"When you go to a teaching hospital you have residents and interns caring for you, which is different from our hospital," said Dr. Richard Zelman, the hospital's director of interventional cardiology, referring to doctors who have not completed medical training. "When you come into Cape Cod Hospital at three in the morning having a heart attack, you have an attending cardiologist with 20 years of experience that will take care of the patient every step of the way."

I'm not confident that the public is aware of the system used for distributing responsibilities for patient care, and more importantly, patient treatment decisions, at teaching hospitals (which tends to be what is meant by "prestigious hospitals" and "prestigious docs"). Both the first-line doctors and the first-line nurses usually are students in training, as are many of the technicians. The professors (the legal attending physicians) are far less a presence after normal hours. In my experience, the perceived ability of the house officer to handle complicated problems rises dramatically after dark. (N.B.: the TV heroes, Ben Casey and Dr. Kildare, were both house officers in training, yet presented to the viewing public as if they were the ultimate docs medicine had to offer.) Many night-time surgical procedures and interventional measures occur without the presence of even a board-eligible doctor, much less a "prestigious doc". The first call from the nurses (who themselves may be in training) almost always goes to the intern or 1st-year resident in pre-fellowship specialties (medicine, ob-gyn, pediatrics, general surgery, urology, orthopedics, etc.), and frequently likewise in the fellowship sub-specialties, or otherwise to the first-year fellow. It is commonly considered a cop-out for this person to call those above him.

Teaching hospitals frequently spend money on complicated equipment that would not be used with enough frequency at a community hospital to justify its cost, and they frequently offer otherwise unavailable teams for managing unusual or exceptionally time-consuming problems. The same is true of much investigational treatment. There can be little comparison with community hospitals in these areas, since the services usually do not overlap.

My guess is that the increased prices at the teaching hospitals reflect their inefficiency and increased overhead. The patient and his insurance company are paying for the education of tomorrow's independent healthcare providers. I don't have a suggestion for how to get around this. In my state, the best results for surgical cardiovascular healthcare, and the lowest expense for producing those results, is provided by a private, for-profit, doctor-owned facility with no teaching or research agenda. But the personnel who provide that care came from teaching facilities such as described.

@burger flipper:

email me or leave a comment for reply on my blog and I'll try to help.

Robin, a "status" explanation sounds like a bias from academia, a world that insiders say is populated by people who focus on status.

Another guess: Brand-name hospitals are less likely to leave a sponge in or take off the wrong leg. And if they do something inept, they are more likely to reach in to deeper pockets to preserve their brand than No-Name Municipal Hospital. Whether all that is true or not is another question. But, it's what the customer would expect, eh?

I like your posts on this topic.

Not sure what is wrong with the market for hospital reputation, but I am guessing whoever figures out how to match patients with hospitals more efficiently is going to be in great demand.

Do you know if Al Roth has looked at your observations?

Silas and Felix, those effects should already be included in the data mentioned.

retired, I'm saying that informing patients of those details won't much change their demand.

Observation: the choices made are strange. They're not what we might have expected.

Explanation 1: Something strange is going on with the consumers - e.g., they're seeking status instead of results.

Explanation 2: Something strange is going on with the sector - e.g., there is massive government interference in medicine.

We already know that Explanation 2 holds. Explanation 1 may be superfluous.

I can imagine the following reply: "I have already taken into account government intervention, and people still are making choices very different from what we might have expected." To which my answer is: I'm not so sure it's that easy to take government intervention into account, to foresee and accommodate all its effects.

The key in the quoted piece is that the mother-to-be "talked to women who had their babies in Boston," that is, her somewhat wealthier friends. I see this in my nabe all the time, although here of course no one admits to going to the hospital - you have to have the celebrity mid-wife and doula who delivered Starlet X's kid by Pretty Boy Y. And you must have a real English nanny from The Upper East Side Nanny Agency.

It's not only that you feel as if you get better care from a higher-status authority, but that the entire child-having process now is as completely status-ridden as buying a car. The child from conception onward is just an object on which to hang your wealth signals via various accessories, from the doula to the stroller to the designer organic Japanese baby clothing to the US$15K kindergarten.


Explanation 1 is a causal explanation. Explanation 2 basically asserts that Distortion A causes Distortion B with no description of the actual causal mechanism. Even if it's correct, it doesn't actually explain anything until more detail is provided.

That Boston article was a bit too conspiracy theorist in tone and overlooked several factors that lessened the credibility of its claims. There was no comparatives of the acuity levels of patients, regardless of whether they're receiving the same procedures. Tertiary centers often get patients of higher acuity than would go to smaller community hospitals -- that alone could credibly explain some of the differences in costs, as well as the "worse" outcomes. But without controlling for acuity, no one can credibly compare hospitals. There were no comparatives for the experience and training levels of staff. There was also no consideration for the percentage of free care provided and larger regional centers also often provide more free care for indigent populations, which raises their overhead for the rest of their patient population. This story didn't provide sufficient detail or balance to allow for any credible conclusions.

No one ever got fired for buying IBM.

It's a subtle distinction, but I don't think the patients want status as an end in itself, just that they are much better at judging status than directly judging quality. They occasionally hear isolated claims that lower status hospitals are better, but they are probably suspicious of concentrated claims. Can you trust it? Can you trust yourself to unpack the details and make sure it is judging exactly the quality you're looking for?

Nurse, did you read any of the studies the paper refers to, or are you just reacting to the newspaper article? Academic studies usually try to control for severity of patient condition.

Constant, your theory that everything government touches is "strange" seems too weak to be of much use. The theory that people seek connections with high status people leads to lots of concrete expectations.

does the only possible arbitrage here involve educating people?

Robin -- have you noticed that observational studies in medicine are often overturned by randomized controlled trials? Its not because the observational studies didn't employ the fanciest techniques in statistics to control for disease stage or acuity --- its because the techniques we have for measuring this stuff just aren't accurate enough.

I know economists believe they can control for confounding -- it just amuses me how they ignore the ridiculous number of well conducted large observational studies that have been overturned. I'm sure you plan on contrasting the level of evidence in Medicine vs. Economics at one point.

Patients might be matched on 5 variables -- but any one of 100 to 500 other features might obviously indicate to any health care professional that one patient is much, much sicker than the other. Anyone who has ever taken care of a person who is sick or talked with someone who has knows this.

The absolute sickest patients are the ones transferred to tertiary care centers. This is the way those hospitals work. It doesn't matter what you control for -- its just gonna be absolutely MEANINGLESS. Garbage in, garbage out.

My one quibble is to question how much status people actually gain by going to certain hospitals. I've heard that some chimpanzees are willing to "pay" to be able to look at high-status chimps. I don't think looking actually raises their status though.

Diogenese, if even the best research can't disentangle these effects, what basis does the average patient have for believing prestigious hospitals are more healthy?

Nazgul, this hypothesis predicts education will have little effect.


"this hypothesis predicts education will have little effect"

Indeed, only increasing the status of the actually safer and cheaper community hospital will work, it seems. Considering the enormous difference in medical prices between the high-status and lower-status hospital, this should turns out to be a strangely urgent piece of health care policy.

How to increase the status of the community hospital - have the city or state give out awards to place on the hospital's website? Have the hospital run ads announcing these awards? Run ads highlighting its most accomplished doctors and telling success stories? Painting the rooms is more fashionable colors and upgrading the decor?

I just put an added on the post.

frelkins, paying all the hospitals the same amount would go a long way; don't expect much support for that though; people like high status folks being paid more.

Robin -- doctors at the academic hospitals -- especially top hospitals -- are often super-super-super specialized. It is economically inefficient, but it lets 1 person become an expert on a particular disease, surgery, or procedure, and to study and try and advance it. If you're a healthy 35 year old or have a common disease, the benefit is probably non-existent to marginal. However if you have a rare disease, that most physicians only see 1 or 2 in their careers, you'd probably be better off going to an academic center where someone has seen 100 patients just like you. Similarly if you have a common disease that has done something peculiar, at an academic center, there will be someone who is familiar with it.

@retired urologist: I think the quality of care you can receive for common interventional procedures is probably equivalent at top community centers (maybe even superior) than at an academic center. However, for elective procedures (lots of cath, angioplasty, CABG, spine surgery, low risk prostate) -- I really, really have to wonder whether the risks and benefits to intervene (and generate $$$) is provided to the patient in the most honest way. Although the outcome is good -- the question of whether or not you even needed the intervention -- and where you'll get an honest answer on that, is a whole other story. At least at some academic centers, there will be people conducting trials on said intervention, and whether or not it helps.

"This is a teaching hospital; you're here to learn. We could call it a learning hospital, but that would scare the patients." -- 'Gideon's Crossing'

Maybe people like going to the expensive, prestige hospitals because they feel like, if things go wrong, they are allowed to complain. Also, the hospital would want to keep these high-paying customers so the customer service may be better.

I'd say the patients are perfectly rational. They don't have good information on quality of healthcare, so they use proxies like prestige and price. As long as there's a non-negligible positive correlation between these proxies and quality, what patients are doing is perfectly rational, especially since there are good reasons to be very risk-averse with healthcare.

Or are you saying the correlation between quality of healthcare and prestige/price as averaged over all healthcare providers is negative or very definitely exactly zero? I seriously doubt that.

Why do you think that going to high-status hospitals despite the facts is caused by desire to associate with high status, rather than stubborn inability to change one's visceral opinion about quality of service in response to evidence? The latter looks more plausible to me, is there a fact to change my mind on that? Nontrivial evidence isn't believed, people may ostensibly agree with it, but make decisions on a cherished gut feeling that wasn't moved.

I confess I haven't read the article, so maybe I missed something. But what justifies your belief here?

Yes patients want prestigious docs and hospitals, and they may believe that high status docs are healthier, but I doubt that belief causes their demand; causation most likely goes the other way.

I think that I as a consumer rely on a general rule to the effect of "more expensive --> better quality", especially in fields like medicine where, relative to other services, I would expect the wealthy to care more about actual effectiveness. Even if I'm wrong to expect that the wealthy think that way, or about the implication "more expensive --> better quality" in general, still this belief would influence my buying choices. In the aggregate, this would add to the demand of more-expensive hospitals. So the additional cost of some hospitals would really be due to a perception that they provide better service, not to my desire to affiliate with high-status docs.

Certainly status could explain why one hospital, rather than another, became more expensive in the first place. But once it is more expensive, I would expect patients' desire for better care to keep it that way, even in the absence of any evidence besides the price that the care there is better.

I'm also skeptical that these studies are really all that accurate.

How much of the data is self-reported by hospitals? What if their reporting procedures are different? Are they controlling for all the relevant variables? What if the people in the more prestigious hospital tend to be sicker than others? I'd bet they tend to be older, since older people tend to be wealthier and more able to afford expensive medicine. How well can they control for the severity of the conditions?

I just can't see really sick people caring about prestige. All the really sick people I've been around didn't. They just wanted to get better, and to get better as quickly as possible.

I agree with Tomasz. There is obviously some correlation between prestige and quality of care, e.g. most people don't go to witch doctors and faith healers. I can believe that this correlation is very low, because people do prefer to associate with others of higher status and are willing to pay for this privilege. What I find much more difficult to accept is that they would risk their own health to do so.

One problem that I've run into myself is that it is very hard for a layman to get a hold of data on health outcomes (for procedures or hospitals).

Constant, your theory that everything government touches is "strange" seems too weak to be of much use.

It is not a theory but a reminder that the economy is not happening in laboratory conditions. Physicists isolate what they're studying very carefully from uncontrolled variables. Economists, understandably, cannot do this, but for this reason they should be mindful of them. I did not notice any manifestations of such mindfulness. On the contrary: a reminder to be mindful has been rejected as not "of much use".

The theory that people seek connections with high status people leads to lots of concrete expectations.

Maybe, maybe not. Look at what people are moving in the direction of to a degree that is not accounted for by your initial model, label it "high status", and voila, you have your "concrete expectation" that people are moving in that direction.

Explanation 2 basically asserts that Distortion A causes Distortion B with no description of the actual causal mechanism. Even if it's correct, it doesn't actually explain anything until more detail is provided.

I wasn't trying to provide a competing explanation, I was warning Mr. Hanson against using the medical sector of all sectors as a test case for his novel ideas. He notices that people don't do what one would normally expect in a regular market: prefer the cheaper over the more expensive (for a given level of quality) and choose the better over the worse (for a given price). But this is precisely the sort of departure one should expect in response to government interference that makes it extremely difficult compare prices and compare quality. It's difficult even to know the price of the procedure one is about to receive. Doctors certainly are not eager to share it with you, and even if they did, you get multiple bills (or your insurance does) from various providers - not just the doctor but the hospital and the anesthesiologists, for instance. You receive one bill, you pay it off, and you think you're done, but then you receive another, and another. You don't know it's all paid off until you stop receiving bills. And the bills are incomprehensible, and some of the items are outrageous - hundreds of dollars just to lie semi-conscious in a room for a couple of hours after surgery, for example. This is nothing like a typical capitalist market sector. Of course it's hard for people to act as a microeconomist would normally expect. And I haven't even mentioned the shielding effect of insurance (which is strongly encouraged by tax policy) on all this. As far as reputation, Beth Israel has a fantastic reputation, thought of very highly (I live in the area), whereas I have no such impression of Mass. General, so the notion that Beth Israel is lower status than Mass. General strikes me as being nothing other than an arbitrarily inserted presupposition to fit with the observed difference in payments. Just find the highest-paid hospital, call it "high status", and voila, you have your explanation.

"One influential researcher found that Beth Israel's overall mortality rate was lower in 2005 than the mortality rates at both the Brigham and Mass. General, but the hospital and its doctors still earn 15 percent to 20 percent less for the same work. "

Does the study control for the complexity of procedures performed or the prior health of the patients being treated? Name brand medical centers often take care of patients with far more complex medical problems than your typical hospital. Lower-tier hospitals even transfer difficult cases into the hands of top-tier centers -- and in many cases hand over complete train-wrecks to top centers in hopes that they'll be able to clean things up. Looking at mortality rates alone is misleading at best.

"State health officials have tried to encourage women like Dahl to reconsider their flight to Boston, pointing out in a 2003 study that community hospitals are generally just as reliable as teaching hospitals for normal births."

The study refers only to "normal" births. Of course, it's impossible to know in advance whether you will have a "normal" birth. That's why you consider going to a well-known hospital. For example, what if your baby needs emergency treatment upon delivery and the community hospital is not adequately staffed to provide it? You generally don't have a heck of a lot of time. Are you really willing to roll the dice on the well-being of your baby?

There's no doubt that one typically has to pay a premium to get treatment at top-tier medical centers (at the same time, these centers are not as much in the habit of denying patient care because of insurance reimbursement issues as a private practice might be). There's also no doubt that name brand medical centers are far from perfect. I can't tell what, if any, premium you should be paying. But to cite some dubious studies that leaves out valuable information is highly misleading.

The comments to this entry are closed.

Less Wrong (sister site)

May 2009

Sun Mon Tue Wed Thu Fri Sat
          1 2
3 4 5 6 7 8 9
10 11 12 13 14 15 16
17 18 19 20 21 22 23
24 25 26 27 28 29 30