Q. What is it costing hospitals to provide care? Malpractice insurance is a major problem, and nothing is being done about it. Implant costs for cardiac and orthopedic are vendor issues. Hospitals have to deal with those rising costs, and none of the payers want to work with the hospitals to take care of that. And what is being thought about making it easier for hospitals and physicians to work together? If you can get those two groups together, you'll find efficiencies that will be unbelievable to providers.
A. Let's start with medical malpractice as an example. I'm going to give you a yes, no, yes answer to medical malpractice's influence on cost. Every medical professional believes that medical malpractice significantly affects cost because of defensive medicine or what have you.
The "no" is the academic literature suggesting that medical malpractice has had a very modest effect on defensive medicine and supplier-induced demand. One study says zero, another study says not estimated, another study says zero. The reason they are finding that is they tend to look at, for example, the variation in cost, and they try to correlate that with the stringency of medical malpractice systems in those areas. If you look across different regions, there's not a high correlation. It's not the case that the higher cost regions have disproportionately stringent medical malpractice systems; therefore, they conclude no relationship.
But I think that misses the point, so I'm going to come back to the "yes." Most medical malpractice systems across the United States don't have massive differences. You're looking at slight differences in the stringency of the medical malpractice system in one area relative to another. A lot of what is happening is practice norms, partly because social interactions between doctors vary across parts of the United States, possibly because of fears of medical malpractice. That is to say, the existence of the medical malpractice system reinforces social norms among doctors because one of the defenses is, "I'm just doing what the guy down the hall did." Similarly, the patient is saying, "My neighbor just got da da da." So it induces more intense practices in some areas than in others even though if the system were slightly less stringent in that area, it wouldn't significantly affect the strength of that social norm. That is to say, the existence of the medical malpractice system itself may induce this kind of variation and inefficiency even though small variations and stringencies across different areas don't have that big of an effect.
Also, if we had a dramatically expanded comparativeness effort and clearer guidelines on when spinal MRIs should be ordered for lower back pain, a defense against medical malpractice would be that we were following the guidelines put forward by whatever entity promulgated them. You might undermine this social norm impact of medical malpractice if much more of the medical system were backed by specific evidence that it worked. That would be the defense. That would become the safe haven. You were following some best-practice guidelines that were determined somehow.
Coming back to physicians in hospitals, what is it about the Mayo Clinic, for example, that generates more efficient outcomes than UCLA Medical Center? Is it that the physician model is different? That sounds promising, and maybe that is indeed the case, but I note two problems or two counterarguments to introduce a little bit of caution.
The first is even across other hospitals within the Mayo system, you see variation. Then perhaps more dramatically, if you look at the VA system, where you think you have the same kind of efficiency benefits, you still see within the VA system a significant amount of variation across the United States. If the primary indicator of inefficiency is this variation that we can't explain, you're still seeing a lot of it in the VA system. I'm sympathetic to the perspective that we need to align physician and hospital incentives much better than we often do. I just think that we need to unpack a little more how to do that.
That raises another point. Over the next couple of years, we should be doing much more aggressive demo projects on payment methodology changes, on different models of providing incentives for doctors, and what have you within the Medicare program that we're not currently utilizing. Most of the demo projects within Medicare aren't particularly well designed to teach us anything, and we could be much more aggressively experimenting. We need to if we are going to tackle this underlying problem. Clearly, incentives for physicians and hospitals are among them. We should be trying lots of different things and seeing if we can unpack some of the puzzles that exist so that you all can hopefully apply them in your own hospitals.
Q. For things like hip implants or total knee replacements, the cost of an implant can vary. Many orthopedists have told me they can use a generic knee for most of their patients. The technology is probably six to eight years old and can cost from $500 to $800 up $10,000 to $12,000 on a reimbursement to the hospital of about $9,200. Is it possible any time in the future to allow a Chevy Impala, if you would, for everyone and allow them to buy up, as opposed to allowing patients to select whatever level they would like with no consequence economically for them, but have the cost passed on to the facility?
A: Sure. Let's even broaden it out. What about beneficiary cost sharing? The evidence suggests beneficiary cost sharing will reduce cost. But the question is how much? You face this tension between two objectives. One is to get beneficiaries to become more cost conscious, and the second is that the purpose of insurance is to protect against catastrophic costs, so you presumably want deep insurance and therefore little cost sharing for catastrophic cost.
The problem is that such a large share of total costs come from that catastrophic area. For example, the top 25 percent of Medicare beneficiaries ranked by cost account for 85 percent of total cost. Unless you're going to apply stringent cost sharing to those kinds of cases, which would then undermine the purpose of insurance, you don't get that much traction from the consumer side or from the beneficiary side. You still get some, not only because some costs are not catastrophic, but also because the evidence from the RAND experiment in the 1970s suggests that when you apply cost sharing to the noncatastrophic costs, you get some reduction in noncatastrophic costs also, because people are less likely to enter the expense of modalities in the first place, basically. Nonetheless, there's not as much traction as one would hope as long as you're going to provide generous insurance against catastrophic cost because of that tension, which is why I put more emphasis on the incentives that we have for providers to provide more efficient care. A lot of the health care delivered in the United States is, especially in the catastrophic cases, basically what the doctor orders. So the question becomes how do we change what the doctor is ordering?
Q. Please give us your perspective about the economic political environment in which you operate and what that means regarding the federal government's participation in Medicaid. In other words, as you have more and more people and more and more interest aligned toward keeping Medicare reimbursements the same, does that same level of people want to keep the federal government versus patients on Medicaid along those lines? Is that an area that you think politically is easier to attack than Medicare?
A. In Medicaid, as opposed to Medicare, we may be closer to the boundary of access problems. It's very likely that Medicaid will be a part of any stimulus bill that is passed. We will have a temporary increase in the so-called FMAP-that is, federal medical assistance percentage, or the federal match rate for Medicaid-which is intended as general fiscal relief to state governments. It is true that money is fungible, but it is also true that the evidence on things like that suggest the so-called flypaper effect, which is that if you provide a subsidy channel through a particular program, even though money is fungible, it's not perfectly fungible. Some of the money sticks to that program, which is to say an increase in the FMAP probably will provide some support to the Medicaid program itself as opposed to just the typical full fungibility assumption that it will simply spill out into overall state and local governments.
By the way, these rising healthcare costs, even though gradual, are already manifesting themselves in lots of different ways. At the state and local government level, the evidence suggests that rising Medicaid costs have crowded out state government support for higher education. And tuition, which is the other main force for public universities' main funding source, has not increased sufficiently to offset the reduction in relative funding from state and local governments. That is not surprising given that 20 years ago state support for higher education was twice as important as tuition. If you reduce that by 20 percent, you'd have to raise tuition by 40 percent to offset it. Not going to happen. The dramatic increases in tuition that you read about they not been sufficient to offset the relative changes in state government funding. The result has been reductions in spending per student and in new assistant professor salaries at public universities relative to private universities.
So, for example, if you look at Berkeley versus Stanford or UT Austin versus Rice, or you can do all the comparisons, public versus private, 20 years ago, new assistant professor salaries were about the same at public and private universities. Now they are about 10, 15, or in some cases, 20 percent lower at the public universities. Again, you can trace that back to rising healthcare costs affecting state and local governments, affecting public university budgets, showing up in assistant professor salaries.
Similarly, one of the reasons that we don't have as much demand for efficiency in the health system as would otherwise be the case is that such a large share of healthcare costs is delivered though an employer-sponsored system, where most workers don't realize that their take-home pay, at least in aggregate, is being reduced to finance those contributions. All of the economic evidence and all of the theory suggest when employers pay $200 or $10,000 for coverage, the result is reduced take-home pay for workers. But we don't process it that way. Think about the backlash that you hear about out-of-pocket spending. Then consider the fact that out-of-pocket spending is only 15 percent of total personal healthcare spending, and imagine the demand for efficiency that would come if we multiply that backlash by seven. I think the demand would be significantly larger, and I think most workers don't realize how much health care is actually costing them, because it's not coming out of their pocket, in a sense.
I'll end with a little vignette to make that point. Two Harvard professors, very well known, were writing a paper about employer-sponsored insurance. I said, "You should write about the degree to which people don't realize how much health care is costing them because of this form of the contribution." One of the professors, an economist, got really excited. He said, "That's a really great idea." Then he realized he didn't know how much Harvard was contributing for his health care. So he spent about a half hour on the Harvard Internet looking for information about how much Harvard contributed for his health care. After that search, he realized that he wasn't on Harvard's plan because he was covered by his wife's plan. His co-author then said, "Don't you realize that Harvard has been providing information on its paper pay stubs like many employers do about what the contribution is? You could have just looked in that box, and it would have said zero." At that point, a Harvard law professor-not one of the co-authors-came up and said, "But don't you realize that Harvard discontinued paper pay stubs three years ago?" Here were these brilliant economists and law professors who didn't know what information Harvard provides and doesn't provide, let alone and more important what the actual contribution for a full professor in the economics department would be.
Replicate that millions and millions of times across the United States. There's very little awareness of how much health care is actually costing us in terms of reduced take-home pay, which then feeds the political economy problem because it results in less demand for efficiency than would otherwise be the case.
Publication Date: Thursday, January 01, 2009