Productivity and Health Care:
The Value of Medical Technology

February 28, 2001

Transcript prepared from a tape recording

12:30 p.m.	Registration
1:00	Panelists:	Michael E. Chernew, University of Michigan
		Mark McClellan, AEI and Stanford University
		Mark Pauly, University of Pennsylvania
		Jack E. Triplett, Brookings Institution
	Moderator:	Newt Gingrich, AEI
3:00	Adjournment

Proceedings:

MR. GINGRICH: This is one of a series of forums we'll be holding to look at both what I think is the coming transformation of health care, and the degree to which the general process of change in the society at large ends up getting reflected in the health system; and to what extent it doesn't get reflected because of the ways in which we've organized the health delivery system.

I'm Newt Gingrich, former Speaker of the House, and a scholar here at AEI. And I'm going to introduce each of the panel members before they talk. But let me just start by saying that I think it's fascinating: There was a recent Wall Street Journal article entitled, "Productivity Gains Extend Beyond Technology Area." The subhead is, "Productivity Gains Seen Spreading Out."

And it basically builds on Alan Greenspan's various descriptions of the degree to which in the last five or six years we've now had a dramatic increase in the application of capital and technology to improving productivity across the whole society.

The interesting question will be: To what degree is that true in health, and to what degree is it not true? And to the degree it's not true, is that a function of health being unique, or just health being uniquely maldesigned?

That is, are there characteristics of health that make it necessary not to be productive, or are there simply characteristics of health that are blocking the kind of productivity increases you're seeing in other parts of the society?

I think as a part of that question, we need also to ask the question: How do you measure value in production? That is, something may be more expensive than its predecessor, but so dramatically more effective that in fact its cost per value delivered is much less. And I think that's an additional factor that we're going to look at today and discuss.

What we hope to do is have a brief introduction by each of the four guest panelists, and then have a chance to toss it open to the audience to allow you to also engage in this.

And I would hope that there would be two sets of focuses. One is the current state of our knowledge. And coming out of the current state of our knowledge, what should we be doing next? Either what should we be doing to gather more knowledge; or what does the current knowledge, if it's adequate, tell us we ought to be doing to modernize the health system and to try to maximize the degree to which we can have productivity gains and value gains in terms of the way the health system operates?

Let me first introduce as our first participant Mark Pauly. Mark currently holds the positions of Bendheim Professor and Chair of the Department of Health Care Systems. He is a professor of health care systems, insurance, and risk management and public policy and management at the Wharton School; and professor of economics in the School of Arts and Sciences at the University of Pennsylvania.

He is a former commissioner on the Physician Payment Review Commission, and an active member of the Institute of Medicine.

One of the nation's leading health economists, Dr. Pauly has made significant contributions to the fields of medical economics and health insurance. His study on the economics of moral hazard was the first to point out how health insurance coverage may affect patients' use of medical services.

Dr. Pauly is an associate editor of the Journal of Health Economics and of the Journal of Risk and Uncertainty. H has served on an Institute of Medicine panel on public accountability for health insurers under Medicare.

Dr. Pauly.

PRESENTATION BY MARK V. PAULY, PH.D.
PROFESSOR, HEALTH CARE SYSTEMS, WHARTON SCHOOL;
PROFESSOR, ECONOMICS, UNIVERSITY OF PENNSYLVANIA

DR. PAULY: Thank you. When that comes up, there are a few cards I'm going to place on the table--or maybe lines in the sand--about this subject. This is kind of an in-joke, but just in case anybody is wondering, I will not be talking about the tax subsidy for employment-based health insurance today.

[Laughter]

DR. PAULY: Instead, I do want to talk about technology and productivity in the medical care sector.

[Slide Presentation, as Follows:]

Okay. So here are a couple of stylized facts that, actually, I know the subsequent panelists will elaborate on. Basically, most of the data that I know about indicates that the lion's share--whatever that share is--of the growth in real medical spending per capita, even after you adjust for the aging of the population, is accounted for by what we call "technology." Although the dirty little secret is that that's really the residual in the statistical analysis, after we adjust for the things that we do know how to measure instead of things that we don't know how to measure.

And probably--and this will be turned to, almost surely, in the remarks Jack Calfee is going to make--probably, this technology has been quality as well as cost increasing. And indeed, a good bit of the cause of the problem of measurement here is, we don't know how to measure that quality.

I will just mention, partly for historic reasons, but maybe because it will come up later, I've taken the view that we ought to measure it by what people are willing to pay for it and get a value measure that way.

But the two policy questions I want to just address in just a few minutes here are: First, this process we've been observing, is it efficient or inefficient; or do we know? And second, why has there not been cost-reducing technology in health care as there has been in many other sectors of the economy?

Okay. Well, I give the answer to question number two: "Because scientists are too smart." By that, I mean there certainly are examples of technologies where the cost has fallen and the quality has improved in health care. But biomedical scientists have been able to discover new treatments and new cures for previously untreatable, and therefore cheap to treat, disease. The new treatments, at least initially, come in at a fairly high price. And because we've had the bad luck of having more of those, and they stay ahead of the cost-reducing parts for the old treatments, that's probably part of the answer to that empirical question.

But mostly, what I want to talk about is the answer to the first question, the normative question about efficiency. And I'm going to do the "on the one hand, on the other hand" one better here: I'm going to do it twice. So I'll say, yes, it is efficient; no, maybe it isn't; yes, it is; no, maybe it isn't. So that'll be kind of the things that I want to talk about here.

Why the cost technology growth is okay: Well, fundamentally, if you did think it was okay, I think an argument would be made that it meets the market test. The health care market is highly competitive now. Health insurers compete; hospitals compete, as we know all too well in Philadelphia. Even physicians compete with each other.

And a new technology offered on the market has to meet some kind of market test. And at least arguably, if the consumers are willing to pay for new technology, or have their employers take the money out of their wages instead of giving them a raise, to pay the extra health insurance premium, it'll pay the new technology. An economist is inclined to say that's okay.

Or the other way to put it: I keep trying to get venture capitalists to invest with me in this health plan I'm thinking of starting. If you don't believe that the technology is okay, this new health plan's advertising slogan will be "Last year's technology at last year's premiums." And so far, I have not gotten any investors for that particular health plan.

[Laughter]

DR. PAULY: Secondly, even timid insurers actually have the power, at least in theory, to refuse to cover experimental technologies that are not proven. And managed care can refuse entirely to cover, or can also limit the use of new technologies. So at least on the surface, it seems there are enough instruments out there available to insurers to refuse to cover technology that may be good, but maybe isn't worth its cost, or may be utterly useless altogether. You can refuse to cover it.

Trouble in paradise, though: Will this work perfectly? There's at least three things that I'll mention here. One is what I call the "product differentiation problem." Doctors argue--and probably, they're right--that, at least given what managed care insurers are paying them these days, they can't be paying attention to each individual insurer's special rules.

So they tend to do what the average plan or insured person wants. Which not only means they may choose the rate of use of new technology that's sort of the median, as opposed to what some particular individual patients might want; but more seriously, there may be less gained from limiting technology in a particular insurance plan by imposing hassle on the patients if the doctor isn't going to go along with the insurer's effort to hold down that cost. So all you end up with then is irritated insured people, but the doctor still ends up prescribing the scan or the drug or whatever it is that the plan would rather not have prescribed.

Secondly, there is the usual whipping boy here. We might as well talk about the legal system. State laws and courts do force insurers to pay for new technology, especially--and that's most problematic, I really meant here--if it's very beneficial, but very high cost. They'll even force that. Cost effectiveness is not usually a legal criterion for failing to cover an effective technology.

And finally, research on cost effectiveness is a public good. And so the kind of knowledge that all might benefit from about how much good do these things really do, and how much really do they add to cost, in ways we can actually believe, are things that oftentimes are not known; even though the social benefit of gathering that information, summed over everybody who would use it, could well be enough to cover its cost.

Solutions: Some of these problems permit of some solution. The solution to the problem of the product differentiation is to have a doctor who contracts with just one insurer, and that's the staff or group model HMO. The Kaiser type managed care plan can get the doctors' attention--the Kaiser Permanente doctors' attention--because their only employer is the Kaiser health plan. They can internalize these decisions about technology.

And also, I would mention the large plans, especially Kaiser, can and do do technology evaluation. And they seem to find that it pays; although I'm not sure the final version of the jury is in.

Maybe rogue lawyers and rogue judges can be controlled by good research. Autologous bone marrow transplants were at one time covered. And the research finally showed, I think--although somebody may differ with me here--that they're not that effective. And now they are no longer covered.

And then I didn't put it on this slide, because it would seem too self-serving, but I thinks grants to health services researchers to have them do good cost-effectiveness studies might also help.

But there are some worries for the future. And this also gets to the question of why costs don't fall in the health care sector per unit like they do elsewhere. Again, it's sort of a reiteration of my original statement about scientists being too smart and too productive.

The new technology often doesn't stay around long enough for cost reduction to be apparent. It gets pushed aside by a yet new technology which is somewhat more expensive. As everybody knows, the Jeep Cherokee--not the Grand Cherokee, but the regular Cherokee--is an old model that they kept on just because it was cheap. That works in autos; but so far, I don't think we've seen it work in health care.

Here's a point which I think is sometimes missed, but I think is important: Who to blame for this? Well, in a way, some of the most important sources of new technology have been patented products when the Government gives a legal monopoly.

And although there are certainly benefits from that in the form of incentives for developing the new technology, there is not just a cost, but a kind of distortion; in that the cost reduction that may be going on behind the scenes at the drug company, at the device manufacturer, doesn't necessarily show up in lower prices to consumers, as long as the product is protected by a patent. It may show up when the product goes off patent, and then the price falls to the floor.

But we don't see--There's the sort of artificial screen that prevents us from seeing, unless we look really hard at what's happening to real resource cost, as opposed to spending.

And one of my comment themes is: Spending in health care isn't the same as true costs, as well. So there is a difference.

And then, the third thing is this is just sort of a concrete version of the problems with the legal regulatory system. I think a lot of people think, "It just seems so obvious. Wouldn't it be better, instead of having a scribbled prescription that I have to wad up and carry in my pocket to my pharmacy, wouldn't it be better for my doctor to be able to transmit my prescription there electronically?"

And there is money going into that from the Internet side and from the capitalist side. But of course, New Jersey is an example of everything here. And at least in New Jersey, you wouldn't be allowed to do that, because the New Jersey laws only permit the doctor [sic] to dispense a prescribed product on the receipt of either a written or faxed prescription. You can't e-mail your prescriptions in New Jersey--or your doctor can't e-mail your prescriptions in New Jersey.

So those things can be impediments. And I think before we get too excited, we probably ought to have a really serious look at what all the impediments might be. There are more than you could ever imagine.

The biggest issue, though, I think, to the implementation of new technology, especially of the information improving sort, in the health care sector is physician resistance. And we had a very spirited discussion at lunch about this. There were no physicians present. Unfortunately, Mark McClellan couldn't be here, and Jack's going to put on his white lab coat and try to fake it here. But he wouldn't have reflected the views of the people we're talking about here anyway.

So here are a few hypotheses I have. I regarded this as an enormous puzzle, why there has been physician resistance to the introduction of what seems like cost and quality improving new technology. But here are a few of them.

Doctors are not self-selected to be productive. At the MBA program at the Wharton School I run, there are some further self-selected doctors who do want to be productive. And we could actually fill the whole entering class with physicians or medical school students if we didn't achieve a certain type of diversity. But the typical physician basically hasn't thought about particularly reducing costs as a primary objective and practice.

A strong sense of entitlement to stability: "I went to medical school. I aced organic chemistry. What more do you want from me? Haven't I paid my dues in life?"

And then, above and beyond these sociological things, it's not all entirely their own self interest--potentially an incentive to sabotage managed care efforts to control. For example, if you were a physician, how would you respond to the message, "Please submit the data electronically, which will allow us to refuse payment to you"? There's a more complex question. That means that doesn't really have a simple answer, but that's sort of how it looks in the first instance.

And finally, we need to do, as usual, a little bit of Government bashing about rules and regulations which often impede the ability to implement new technology, especially privacy rules, protection against suits if data was lost and so forth.

And finally, little incentive for the physician or the patient to want to worry about cost, if the insurer is going to pay the same amount regardless.

So fundamentally, at least an economist would be inclined to say it's the absence of incentives, and it's the absence of an adequately and effectively competitive market structure.

So here are my last three thoughts. Conclusion: Well, I personally don't get too excited about this. It probably didn't help that the official invitation to this panel was sent to me by e-mail, and on that day the Wharton School e-mail completely self-destructed. So I'm not a big fan of all electronic means.

But I think, arguably anyway, markets are working pretty well, when we let them. And at least the growth of health care spending, which represents higher cost but also better care and better health outcome, may actually be efficient compared to last year's technology at last year's premium.

Are we willing to make the tradeoffs that new technology might imply? For example, what would you think about the notion of, you know, "A special on colonoscopy this week if you book it over the Internet"? Do you really think that's how the health care sector could work? Maybe it would. It's an interesting idea. But maybe we'd have some apprehension.

It's not obvious to me there is need for imposed industrial policy in the private sector. I think there's probably need for better incentives and elimination, or at least relaxation, of rules that prevent innovation. But I believe Medicare could show by example, if it were given looser controls--I also need to say, more money for the administrative part--and maybe enough rope. Maybe that would be the place to start, with implementing some of these systems. Thank you.

MR. GINGRICH: Thank you very much. I can't help but observe that New Jersey, which not only makes it illegal to have an e-mail version of your prescription, also is the only state left in which you cannot have self-service gas.

[Laughter]

MR. GINGRICH: A particular conspiracy which they manage in part by ensuring that the price of gasoline in New Jersey is consistently below New York and Pennsylvania. So consumers have no vested economic interest in breaking the local gas station control system. But it's interesting, the ways in which New Jersey is an outlier in terms of the evolution of new technologies in America.

We were going to have next a briefing by Mark McClellan, who has been a visiting scholar here at AEI. But in what has been a pattern now for the last two months, we keep losing presenters. Because he's now going down to the White House to be the advisor to the economic program with Larry Lindsey. Mark will be focusing on health policy. So he couldn't be here today.

However, he cheerfully sent his slides as sort of a "Farewell, and I'll see you later" note. And we promptly turned to Jack Calfee. And he's been very generous in agreeing to join us to talk with us.

It's very appropriate, because he's been a resident scholar here at AEI since 1995. He's an economist. His PhD is from the University of California at Berkeley. After serving in the Bureau of Economics of the Federal Trade Commission, he taught marketing and consumer behavior in the business schools of the University of Maryland and at Boston University, and has spent a year at the Brookings Institution.

Jack's research is focused on the role of advertising, information, and regulation, especially in health care, on tort liability and related areas. In addition to academic journal articles, he is the author of Prices, Markets, and the Pharmaceutical Revolution and Fair Persuasion: A New Perspective on Advertising and Regulation.

We really appreciate your stepping in, and look forward to your presentation.

PRESENTATION BY JACK CALFEE, PH.D.
RESIDENT SCHOLAR, THE AMERICAN ENTERPRISE INSTITUTE

DR. CALFEE: Well, thank you, Newt. You may appreciate it a lot less when I'm done.

[Laughter]

DR. CALFEE: But I understand that Mark was working under a deadline, and he had to reform Medicare by the close of business tomorrow. And he was not able to take out the time to come over here.

Actually, I may not use those slides, because I'm afraid that I'll get confused if I try to follow the slides and my notes that I've tried to take from looking at materials, because Mark also wrote a paper for this conference which has not been distributed, as far as I know. But it's within the theme of this conference, which is productivity and the health care sector.

And obviously, what I'm presenting are the views of Mark and his co-author. This is not an area in which I've done any original research. I've only kept tabs on it rather at a distance; whereas Mark has been very much at the forefront of assessing what's happened with productivity in health care.

And essentially, what he describes is there have been few approaches to try to figure out whether productivity in this particular economic sector has been moving apace the way it has in some of the other sectors.

And one of these approaches is something that he refers to as a direct method. I think of it as sort of a macro method. And that's where you look at overall outcomes, health care outcomes, and you look at the expenditures in that area.

And the two areas that have been studied rather intensively are--Heart attacks is one of those areas in particular. And essentially, what they've found is that if you go back over 20 or 30 or 40 years, you find that we're spending more money in dealing with heart attacks over the years, but we're seeing tremendous improvements in the outcomes from heart attacks. Mortality rates are drastically reduced from what they were 20 or 30 or 40 years ago.

It's not always easy to tease out what it is that's causing the reductions--and let me flip ahead here--to tease out what's causing the improvements. But the evidence that's been put together so far strongly suggests that the bulk of these improvements are due to improvements in medical technology itself.

The bottom line that comes out of that are two things: One, we're spending much more in treating heart attacks than we used to. We're getting much better outcomes. And then, of course, you've got to move to the third stage, which is trying to assign some value to the improvements in outcomes, and compare that to the extra cost.

And if you do that, using reasonable measures--in fact, a variety, a quite wide variance of measures--the bottom line you get is that we're getting a lot of benefit out of the increased cost. In other words, by any reasonable standard, the benefits from improvements in heart attack mortality greatly exceed the costs of dealing with those things. And so, in that respect it looks like we're getting some very good developments. We're getting much productivity.

Another area that Mark describes in some detail is quite different, which is looking at low birth-weight infants. And again, what we're seeing is substantial investments in medical technology, a lot of new techniques. We're getting substantial improvement. And again, if you try to balance the costs and the benefits, you're getting large net benefits. Which is another way of saying that we're saving lives; we're improving longevity; and by any reasonable assessment of the value of the saved lives, we're coming out ahead. We're getting much greater benefits than we are paying costs.

And that's one way to look at productivity in health care. Another way is what you might think of as the indirect or micro assessment, where you actually look at clinical data. And you follow a cohort of people who have encountered certain kinds of illnesses, and you see what's happened with that cohort. How many lives are being saved? What proportion of lives are being saved? How much are we spending on these particular incidents?

And again, you do this disease by disease. And instead of looking at data that sweeps across the economy, you need to gather clinical data to assess things.

Again, Mark's paper looks at two areas. One of them is depression, where treatment has gone through a fairly fundamental shift in the last, I guess, 15, 20 years or so; where we've seen more drug therapy as a substitute, a partial substitute, for counseling or mental health care.

I'm sensitive on this topic, because my wife is a psychologist who also happens to be a strong believer in drug therapy for depression.

And what the data show is that, yes, we've seen a shift from one thing to another. In other words, we've seen an adoption of what you might think of as medical technology.

In this case, the costs have probably gone down. You save a lot of money with drug therapy, as compared to the mental health therapy. And the outcomes are getting better. So this is one of those situations where costs are going down, outcomes are improving and, again, the net benefits are strongly positive.

And then he also looks at another area where you take this indirect or micro approach, and that's for cataract surgery. And what they find is more or less the same pattern. That is that the costs of dealing with cataracts now are actually substantially less than they were 20 or 30 or 40 years ago. Instead of a person being hospitalized for a fairly extended period of time, often I understand today it's more like 24 hours, 48 hours, for cataract surgery. And so again we're seeing cost going down, and we're seeing the outcome improving very substantially.

And so these studies have results that are very encouraging. And that is that we're seeing what we really like to see, which is improved outcomes combined with lower costs, just as we see in a lot of other sectors of the economy.

Now, there are some down sides to both kinds of research. There's a lot of things that we don't know that we would like to know. For example, going back to the macro or the direct approach, where you look at economy-wide data, we see improvements in outcomes for heart attacks, etcetera; but we don't really know for sure that that's the result of improved technology. To some extent, it could be a result of healthier lifestyles and some other factors. But again, the data strongly suggest that technology accounts for the bulk of those improvements.

When you look at the micro data, the indirect data where you follow individual clinical cases, the results are very enticing. But again, there's a lot that we don't know that we would like to know. For example, the health problems that have been studied are ones that, while they are important, are relatively small compared to some other things. In particular, we haven't seen much research on some of the really big-buck items, such as obesity, diabetes, and other chronic conditions, where you can help, but you don't necessarily have a cure.

And then, there are some areas where we're not too sure there's been much progress at all. And the most outstanding area there, the most obvious area, is cancer, where there is a lot of evidence. And there are a lot of people who have suggested on the basis of what you might call, not anecdotal evidence, but not compelling economic evidence either, some suggestions that maybe we're not really making much progress in cancer. That is, that we're not getting much improvement in outcomes; we're spending more money; and the benefits could even be negative.

And that's basically the second half of Mark's paper, and what his slides would have been if I'd had the courage to go through them. And that is a study that he and David Cutler [ph] have done, an original study, looking at breast cancer treatment.

And essentially, what they find is that the incidence of breast cancer has been stable, or even going up to some extent. And this is probably because of earlier and better detection, rather than some basic changes in health. Mortality from breast cancer has been going down fairly strongly. To some extent, this is due to simply finding breast cancer sooner; which means if you measure mortality over a given time period, if you're detecting the illness earlier, the mortality data is going to look better even if you aren't having any genuine improvement.

But there's also data that indicates that we're getting some genuine improvements. That is, that some lives are being saved that otherwise would not have been saved. The technology is expensive, and so we're spending more rather than less than we used to spend dealing with breast cancer.

But if you put it all together, when you make reasonable valuations or assess reasonable valuations to the value of lives saved, adjusting for the quality of life after therapy, what the data indicate is that we're spending more on dealing with breast cancer; mortality is improving. If you balance the two, it appears that we're still coming out ahead. And that is that we're getting net benefits from the increased expenditure on breast cancer. And that's good; although these net benefits are not as spectacular as they are in certain areas, such as heart disease.

And the bottom line seems to be that that's simply because medical technology in dealing with breast cancer, and cancer generally, has not been advancing as rapidly as it has in dealing with heart disease.

So that's my overall take on what Mark has written. It's unfortunate that he couldn't be here, but that gives you at least the flavor of what he's talking about.

MR. GINGRICH: Well, first of all, let me just thank you for pinch-hitting. Let me also say that many are interested that the head of the National Cancer Institute did a presentation at the Bios [ph] conference in New York last week which at least had some significant hope that in fact on the whole range of cancers we're about to see very dramatic breakthroughs, largely because the spinoffs of the Human Genome Project and the degree to which we're beginning to understand vectors are so dramatic and on such a massive scale that it's very likely we'll gain more ground in the next five years than we gained in the last 25 in terms of understanding and beginning to apply technologies.

And it made Mark's charts look very simple. The stuff he put up was just an immense amount of data, all of it at a surface level--all of which by the way, is available if you go to the National Cancer Institute's website. I mean, they have a continuous updating across the entire range of all of their research, which is pretty phenomenal and very, very powerful.

But thank you again. Pinch-hitting like that is never an easy thing, and we're very grateful.

Let me say that we're also very grateful to Michael Chernew for coming and joining us. He has served on a panel, the Technical Review Panel of the Medicare Trustee's Report. He is the associate professor at the University of Michigan in the Departments of Health Management and Policy, Internal Medicine, and Economics.

Dr. Chernew's research focuses on assessing the impact of managed care on the health care marketplace, with an emphasis on examining the impact of managed care on health care cost growth and on the use of medical technology. His research also examines the determinants of health plan choice, with an emphasis on the role of quality in influencing plan choice.

As I said, he served on the Technical Review Panel of the Medicare Trustee's Report. He is a faculty research fellow of the National Bureau of Economic Research. He is on the editorial boards of "Health Services Research," "Medical Care Research and Review," and "The American Journal of Managed Care." His Ph.D. is from Stanford.

And I think for the purposes of today's program he is particularly helpful because of the report that they issued, and the level of controversy it engendered.

We appreciate your being here.

PRESENTATION BY MICHAEL CHERNEW, PH.D.
ASSOCIATE PROFESSOR, HEALTH MANAGEMENT & POLICY,
INTERNAL MEDICINE, AND ECONOMICS,
UNIVERSITY OF MICHIGAN

DR. CHERNEW: Great. It's wonderful to be here, and I'm going to try and be brief, so I can hear more of your questions.

As was mentioned, at the end of last year I had the incredible pleasure to serve on a technical panel for the Health Care Financing Administration. And the charge of the panel was to review the assumptions that the Office of the Actuary makes regarding the balances of the Medicare Trust Funds.

And key amongst those assumptions was the rate of health care cost growth per capita, age adjusted, for 75 years. And, true, that's also a joke.

[Laughter]

DR. CHERNEW: Just so you know, the current assumption that was being used was that the per capita, age adjusted cost growth was going to be at the same rate as real GDP growth; despite the fact that there was very little empirical support for that assumption. And I was heavily involved, and sort of explicitly asked, "What do you think of that assumption?"

So I'm mostly going to review some of the things that the panel did and said, in light of this discussion. The panel broadly had the consensus that technology, broadly defined--not just equipment, but learning how to do new things or old things differently; but a broad definition of "technology"--was responsible for increases in health care expenditures, often called health care costs.

But I really think it's crucial to draw the distinction between increasing expenditures in aggregate, versus per-unit health care costs. It's entirely possible that per-unit health care costs can be going down because of technology, and expenditures in the health care system as a whole can be rising, often quite rapidly.

And I understand that that's a semantic distinction, but often people ask, "How come we're not spending less because of technology?" That's different than saying, "Technology isn't driving down per-unit costs."

There are several ways in which I think the report can be criticized, and I'll talk about two of them--somewhat, perhaps, unfairly. The first one is, people often say, "So what? We're spending more. What about technology?" And I want to emphasize, the panel doesn't think technology is bad. The panel largely thinks technology is good. I don't want last year's health plan at last year's prices. And I have this bad back, if there's a doctor in the house, and I'm looking for some medical technology.

But whatever the case may be, it's not necessarily bad to have medical technology and have costs go up, if you're getting what you want for that. And it's important to draw the distinction between this type of productivity on average, versus at the margin.

It's also possible that on average we want this new technology, and on average we're getting the outcomes that we want for that new technology, but there are a lot of people at the margin who are using services which aren't giving them the benefit that you would want them to have, given what you've spent.

And it's a lot harder to control technology at the margin, to draw those lines of who gets the technology, than it is to say, "No one can get access to this or to that." And most of the times when one looks at specific technologies, one would never say we shouldn't let anybody have any of the technologies that Mark, for example, looks at--revascularization [ph], bypass surgery. The question is often at the margin. Who gets it? What's that productivity at the margin? And that's a very hard question to answer, and a very important one.

And so, the point that I want to say again is that the panel didn't think technology was bad. The panel thought technology is good.

The second point that I want to make is, a lot of the technological growth that we've seen may have been induced by the health care financing system that we've had. And how we think about that actually turns out to be very important. And we had a lot of discussion about that. And I'll save my comments on that point, actually, until later, if asked.

But I do think it's important to recognize there's a portion of technology that sort of inexorably rises, and a portion that perhaps might be induced by the system, and that portion matters.

Anyway, the bottom line from this whole endeavor--and I'll say this somewhat with a smile--was that the panel recommended that the actuaries assume that health care costs, per capita, age adjusted, grow at a rate 1 percent faster than GDP.

I will tell you, you shouldn't take that number too seriously, for a variety of reasons. Looking out 75 years in the future is hard. There's a bit of folly in the entire thing.

We had this discussion: Imagine it was 1925, and you were trying to predict the health care system in the year 2000. So you'd miss some things.

[Laughter]

DR. CHERNEW: You'd miss the Depression, World War II. Almost all of modern medicine you'd miss. If you read, you know, Paul Starr's [ph] work, Rosemary Stevens' [ph] work--you'd miss a lot of stuff. You wouldn't know. Who knew they'd be able to stop your heart, repair it, and start it again? Who knew they'd be able to stick a trocar into your abdomen and suck your gall bladder out through a straw? Who knew?

And so I'm really not going to say much, and the panel wouldn't say much, about what's going to happen with the genome project and all of this excitement 75 years from now. I'm waiting for "Star Trek" medicine. It's just hard to do.

But the one thing that is clear is that historically people have died. And they never like that. And they wanted medicine to prevent that. And they wanted to feel better before they died. And different types of technologies were applied to help them feel better and die later--with some benefit in that regard.

If you look at Ann Skitofsky's [ph] work, at some point in time that was little-ticket things; other points in time, that was big-ticket things. Recently, it's been pharmaceuticals. I read in the future it's going to be diagnostic technologies. It always seems to be something different. People still seem to die, but maybe later, and maybe not as ill.

And so the panel thought that, all things equal, 1 percent above GDP--which you can barely find a period which was less than that. It was a reasonable number; albeit with a lot of uncertainty.

Now, it's easy to criticize the actuaries for their old assumption, given the lack of empirical support. And one of the arguments that they've used loosely was an issue of sustainability. And it is true mathematically that health care cost growth can't exceed GDP growth indefinitely. You don't need to think about that at any great length. That is actually true. Add [inaudible] and it is still true.

And so the argument was: Since it can't last forever, if we make them grow at the same rate, we won't have this problem that health care costs will continue to be a greater and greater share of GDP. And it turns out that if you run some relatively complicated general equilibrium models, which were run for us by some folks at the University of Maryland, the economy explodes under some scenarios with these types of assumptions.

And it turns out to be true that the key assumption as to whether or not you think health care cost growth of historical levels spells disaster for our society has to do with what you think about the productivity in the health care sector.

And the reason is because what matters a lot is how many people this growing health care sector sucks into it in terms of work. And that's related to productivity.

And when I teach this in class, actually, the cite that I use is a cite from Mark Pauly, actually published in "Health Affairs," talking about the opportunity cost of health care and emphasizing that what matters is not traditional measures of productivity in this regard; what matters is really how many people you're bringing into the system in order to produce a certain level of outcome.

And the reason traditional measures have a hard time figuring this out is because the traditional measures of productivity try and take price growth out of the expenditure number, and that requires measuring prices. And measuring prices in the health care sector is very hard.

What the panel did was, basically, cite Jack and give up. And maybe Jack will talk a little bit about that. But it basically was of the view that health care prices are extraordinarily hard to measure, and historically have not been measured very well.

And if you look historically at the rate of growth in employment in the health care sector, compared to the rate of growth of the per share of GDP devoted to the health care sector, you get a lot more soothing sense of the sustainability of health care cost growth of the magnitude we would suggest, than if you try to use traditional productivity measures and run them through these general equilibrium models. And I think that actually, if you read some of Mark's work, you can see why that is.

And so much of what we did when we talked about productivity wasn't directly related to the important issues it raised in Mark's and David's paper about how much health we're getting, because we're always going to want more health, and there will almost always be room for improvement in health. And how much we pay for that health matters a phenomenal amount in a whole variety of ways. And how much we're willing to devote of our increasing incomes to getting that health is incredibly important.

But I think in the near term the trick is, whatever we decide in terms of what we want to devote to medical technology and the health care sector, we should try to do that in a system that ideally at the margin is giving people the technology that they want and they're willing to pay for. And technology creates this dilemma that that's hard to do, but that's sort of the challenge for the future. Thank you.

MR. GINGRICH: Let me thank you, first of all, for a very refreshing level of candor. You know, in one of my other hats I'm a commentator on Fox. And we've been going through these discussions about, "Can you project out ten years?" relating to another topic that's been big in this city recently.

[Laughter]

MR. GINGRICH: And to have somebody going to cheerfully explain, "We were projecting 75 years, which is hopeless," is actually very encouraging as a way of starting with where the Dow is going.

I also remember when I was first teaching, I was teaching a course in environmental studies in the early '70s. And the hot book at that time was The Limits to Growth, which had a whole series of projections, all of which were exactly right if you took the assumptions. It turned out later on the assumptions were not right, and therefore the whole rest of the book didn't work.

But it's a useful thing to go back and look at, for example, their projection of how much fossil fuel there was available on the planet; and therefore, given this limit, the morning you'd have the last automobile. They were off by at least a century and a half, but it was an interesting model.

And I thought of that in terms of your point--which I think is really a very healthy idea, since I'm an historian by background, not an economist--which is to go back to some point like 1925, and try a similar projection.

For example, you'd have no notion of what the cost of polio would be in the year 2000, because you would assume some kind of patterning. 1925 was actually two years after Franklin Delano Roosevelt got polio. And the cost--because you would assume that people would live longer, you'd have better iron lungs, you would eventually get to an aluminum lung, you might even have a portable lung by 2000.

So you would spend an enormous--This is a tiny sector, but I think it illustrates the point that you can get what Christiansen [ph] describes as a disruptive technology, as opposed to a sustaining technology. And in a way, the Salk vaccine was a disruptive technology. It eliminated an entire class of expenses.

And I think it's a very, very interesting concept to think about. How could we measure current government efforts to analyze things by going back to a different point in time and saying, "Let's apply these assumptions, and now let's look at the current world and see what kind of outcome there is"? I think that was a very, very helpful presentation.

And now, you've already been built up, Jack. We're going to turn to you.

Jack Triplett is a Visiting Fellow at the Brookings Institution, where his areas of research include productivity in health, finance, and other service industries, with a focus on developing improved measures of output for those notably difficult to measure sectors of the economy.

He serves as a consultant to international organizations and to the statistical agencies of a number of countries on issues of economic measurement and economic statistics.

From 1985 to 1997, he was chief economist, U.S. Bureau of Economic Analysis. From 1971 to 1985, he held positions at the U.S. Bureau of Labor Statistics, including associate commissioner for research and evaluation, and chief of the price research division. In 1979, he was assistant director for price monitoring at the Council on Wage and Price Stability.

Before his government positions, he taught economics at Washington University in Saint Louis, and the University of Oregon, where he was also assistant director of the Institute of Labor and Industrial Relations.

He is the editor of Fifty Years of Economic Measurement, with Ernst Berndt, and The Measurement of Labor Cost, and of Measuring the Prices of Medical Treatments.

He holds A.B., M.A., and Ph.D. degrees from the University of California at Berkeley.

Thank you very much for coming here to be with us today.

PRESENTATION BY JACK E. TRIPLETT, PH.D.
VISITING FELLOW, THE BROOKINGS INSTITUTION

DR. TRIPLETT: Well, thank you very much. I am presentationally obsolete this morning, because I brought along my current technology, which is to use overhead slides. For a program on productivity, this is a little ironic.

[Laughter]

DR. TRIPLETT: And I found out that the technology here was one step forward, which is "Power Point," which I can't access. So I'm going to go backward in time and use older technology, which is just to read you this table. In fact, it's a very brief table. So the old technology is probably just about as good as the "Power Point."

I'm just going to read what the available data show about labor productivity in the health care sector. And there's just two numbers that really matter: From 1987 to 1995, the annual rate of change of labor productivity in the health care sector was minus 0.58. That is to say, it fell half a percentage point a year.

Newt, you made a reference to Alan Greenspan's discussion of productivity. And we've had an acceleration of productivity since 1995 in the economy. How did that affect the medical sector? Well, it did indeed improve after 1995: 0.58 decline before '95; an increase of 0.06, that's six-hundredths of 1 percent per year, after '95. So even in the medical sector we had an acceleration of productivity.

But I think the real message of this is, these are pretty low numbers, by anybody's accounting. Well, what do we know about those numbers?

Well, I think what we know about them is they're probably wrong.

[Laughter]

DR. TRIPLETT: In fact, one of the reasons we think they're probably wrong are exactly the list of studies that the combination of McClellan-Calfee told you about earlier; that when we look at the details of what's happening to expenditures--it's a small number of studies; I want to emphasize that--details of what's happening to expenditures on some diseases, what's happening to outcomes, and make some reasonable estimates about what that means for these kinds of statistics, those don't look like negative productivity numbers at all.

So we know enough to know the numbers are probably not right. And the question is, why are they not right? Well, I have to emphasize that all these numbers are put together by producing price statistics.

Now, I'm not going to talk about the details of price statistics, because that's generally regarded as a dull topic to anybody who isn't an index number maven. So I don't want to explore the details. But the point here is that in doing a price statistic for health, or doing productivity for health, one of the things you have to know about are valuations that you put on changes in health care status.

McClellan-Calfee referred to the cataract study, so let me just use that one as an example. It used to be, for a cataract operation you went into the hospital for some prolonged period. I've forgotten what it was--a week or five days or something like that. You had to be immobilized, because if you weren't immobilized the new attachment sort of duct dislodged itself. So you couldn't move for this period of time. Then there was a lengthy period I think you were also in a darkened room, or something like this. So you couldn't watch television. You had to sort of just lie there.

Now, cataract surgery is sometimes an out-patient treatment. So we've had tremendous change in the technology of doing this operation. But the costs have also gone down. Now, how are you going to handle that if you're going to do a price measure?

Well, suppose--I'll just make up some numbers--suppose the five-day stay in the hospital in the old method costs $50,000, and the outpatient treatment costs ten. I have no idea whether these numbers are off by an order of magnitude. I've just chosen them for an example. We know that the new treatment is cheaper, so let's say it costs $10,000.

Get back to Mark's point about: What will you pay? Well, clearly, if you had the choice between the new procedure, which lets you go out and sort of function fairly normally rather soon, or spending five days in the hospital immobilized--

[Tape Change]

DR. TRIPLETT: --for the cheaper one, if you were asked that question. If you had to choose, if you had to go back to the days in which you were immobilized, and someone said, "Well, we'll give you a choice: $50,000, plus five days in the hospital immobilized, or let's say--" You know, this guy is going to make a little money, right? And he knows it only costs ten, but you don't know that. So he says, "I'm going to give you a choice. How about $60,000, and you can go home that day?" Most people would probably say that that $60,000 operation is a better deal.

Well, that's exactly the kind of valuation you need to get to do these numbers right. Because if you just took the ratio of the costs, you say, well, I'll take the $50,000 operation, and then I'll compare that to the $10,000 operation. They're the same thing; they treat the cataract surgery. So the price fall must have been, you know, one-fifth--80-percent price fall.

But that's assuming that the two operations are really the same. So we've lost the benefit of the improved technology. We've just simply looked at the change in cost. That's meaningful, but it's not all of the whole story.

You need to multiply this kind of anecdote for one particular eye operation and one particular disease to a lot of different diseases, to figure out what that productivity number should be. Because the problem here is that measuring the rate of inflation in medical care is exactly the same thing as measuring the rate of productivity in medical care. If you can do one, you can do the other.

We still have this notion--I think that it's, what, 1993 or so when this notion got put into a big legislative proposal. We still have this notion that the major problem in medical care is runaway inflation. The prices as measured in medical care have risen more rapidly than the CPI for a very long period of time. But the difficulty is that the prices as measured don't have any way of accounting for the real gains that you get out of that enhanced operation.

If we could put a value on the improved operation for the price measure, for the inflation measure, we'd also know how to value it for the productivity measure. So we'd count that new operation as being better, more medical care than the old operation.

That's really the problem we face in even understanding whether or not there is productivity change in the medical care sector. It's this difficulty of trying to figure out what's happened to the technology, first. Secondly, what's the medical care outcome that that technology has produced? That's difficult in itself. And thirdly, how are you going to put a value on the changes in medical care outcomes?

Now, the topics that McClellan talked about, those studies, that's about all of them. There are one or two that I could think of that weren't mentioned. That's about all the studies that really exist currently. That's a small number of studies.

They're very suggestive. They're very valuable. I would like to raise one reservation--two reservations about those studies, that Mark might well have in his paper; but if he doesn't, I'll raise them anyway.

One is that this research is not a random event. We don't go out and do research on these topics by simply drawing the subject out of an urn. Why do people study heart attacks? Well, I've talked with David Cutler [ph] about that. Sure, they knew there was a lot of technology going on. That was a very interesting thing to study.

The same thing is true of some of the other studies that were cited in this. We study the ones where we think there's the most action, in terms of technology and in terms of benefits.

There was actually a study done on arthritis which didn't find much effect, because it's a really hard thing to study. These studies are not a random sample of medical procedures, so they may give us a bit of a bias about how big the productivity improvements are. If you generalized across all medical procedures, you might get an estimate which is biased in the opposite direction of what we now have.

Well, I tried to do this in one case: Take the existing study; make some heroic assumptions about what's going on to other related fields; and then, you know, generalize always, just to see what the outlying numbers were. I got about a 5-1/2 percent improvement in productivity in the medical care sector.

Now, that's not a good number, because it's based on, as I said, extrapolating a lot of studies in ways which probably are not valid for a lot of different reasons. But the point was that neither is the number that showed the half percentage point decline over this period a good number. Somewhere there's a big range of possible outcomes of the productivity figures. One hopes that the true number is somewhere in the middle. But at this point, I'm just not quite sure what we can say about the outcome of these measures.

And the second small reservation about it is that the research that we've got on this topic is mainly using the outcomes of clinical trials. Clinical trials don't always get translated into the same effectiveness when you get into actual practice.

And to bring up a point that was also kicked around in our lunch, there's nothing built into this clinical trial information about the rate of errors in application of technology in the hospital. So that would also produce an upward bias to these productivity estimates based on research.

But nevertheless, my concluding remark is to repeat what I said before. What we know about the aggregate productivity numbers for medical care is that they aren't very good, and we don't know enough to know exactly how to get a good measurement at this time.

MR. GINGRICH: We're about to turn this over to all of you. Thank you very much, Jack. Let me make three observations before we do.

First, I think that, for me at least, one of the most important outcomes of talking with folks about the data on health is the degree to which we really need an effort to define a baseline.

I mean, the fact is, every time you talk to anybody who studies this area, they'll tell you there's virtually no good data; there are virtually no systems for gathering good data. And next year, if we have a conference, the same will be true.

And it seems to me--again, biased by my background, having served on the Aviation Subcommittee and having looked at a system where we took safety seriously and where we actually had lots of data about lots of things--that it is amazing this late in the information age, the paucity of routinized information which could be gathered in a system in which the Federal Government pays over half the cost.

And so it does seem to me that there are a variety of steps that ought to be taken. And maybe one of the outcomes coming out of today ought to be to think seriously about, for the Bush Administration, what is the data we should expect that should be gathered on a routine basis, all of it electronic, all of it done nationwide, and all of it done so you could get to an intelligent decision system about what's happening, what needs to be happening, and how we optimize health in the country?

Second, I happened to get, as an example of a free-market impact, a report that was done from "Refractive Market Perspectives," which is a newsletter on the refractive surgical market, which simply pointed out that we've had a dramatic increase in laser surgery. Laser surgery is a free-market system. This is for eyes. And that because it's a free-market system, it's actually directly competitive.

Prices have consistently come down. We are now in sort of a next-generation laser procedure, and the result has been a continuing decline in cost; a continuing increase in availability; advertising that emphasizes both quality and cost; and a marketplace which has migrated to 1,300,000 procedures a year by people who look around and say, "Who's doing it? What's their track record? What do they charge? And is this something I want to do?"

And I would just suggest, if you look at the substantial decline in the cost of eye laser surgery--in one of the few areas of health that is both sophisticated and free market--and were to compare it with other areas of health, you would find it to raise a very interesting set of anomalies.

Lastly, I couldn't help but think, after Mark Pauly's comments and the whole notion about offering a firm that would give you last year's health technology at last year's premium, it's fair to say, I think, at the present time that, at least in terms of technologies, Medicare tends to give you five-year-ago technology at this year's premium.

And I do think there's an interesting anomaly there, that we tolerate for senior citizens a very long lag time. After the FDA has already taken a long time to approve something, there's then a second lag before seniors have access to whatever it is the FDA has already approved. I would suggest to you that the politician who ran on the promise, "Your parents will get things five years after you," would probably have a little bit of a difficulty carrying that out as a successful campaign promise.

What I'd like to do is just toss it open to the audience and allow you to ask questions. Yes, sir?

[NOTE: Technical problems with recording at this point.]

PARTICIPANT [In Audience]: This is a question for Mike Chernew [inaudible]. With respect to the exercise [inaudible]. The question is, why do we do this? And the reason of course is [inaudible]. You want the estimates at each point in time to be as accurate as possible. Part of your remarks I think directed at your previous method projecting into the future [inaudible] seemed in some sense to be more than I think they should be.

The second consideration, though, is how much weight we've got to attach to numbers that have successively more [inaudible]. Another part of your comments, and I think the papers do reinforce this, was the observation that if you, with this backwards historical exercise, ask how much we would do about [inaudible] in 1925 we would say [inaudible]. The idea that projections made in '25 [inaudible] could influence decisions made in '95 is not clearly pushing us in the direction of rational [inaudible].

So my question is, are you confident that an arguably more accurate numerical projection will improve decision-making [inaudible]? Wouldn't you feel better if, along with your change in the numbers plugged into the equation, you had a way of discounting the weight attached to those numbers?

DR. CHERNEW: Right. Thank you very much. I agree very much with the spirit of your comment.

First, let me say something that I want to be very clear on. Another conclusion of the panel was that the work done by the actuaries at the Health Care Financing Administration was exemplary. I mean, we were just astounded at how thoughtful and how good a job the whole group of people involved in that process did, both in terms of their desire to get the best numbers possible and their dedication to doing that, and their understanding of these issues.

One of the other reasons why you could justify the assumption of the same growth rate as GDP is it allows the forecast to focus only on issues of aging and demographics, and takes out of the equation issues of health care cost growth, the part that's obviously most uncertain.

Your specific question was about how much weight to put on the numbers at the end. And I will tell you there's another chapter of the report that focuses both on how to deal with uncertainty--you know, how to deal with the fact that there's uncertainty, and that uncertainty grows over time--and then, how to deal with presentation issues, how to present these numbers.

One of the things that I think is clearly true is, there is sometimes a tension between presenting in a thoughtful, as accurate as possible manner some information you think to be true, and then conveying that to folks who will then use that in the way that you would want them to use that.

Actually, the panel was quite sensitive to the different ways--you know, the short-run period, the intermediate run, the long-run period--and sensitive to differences in levels of accuracy in projecting over those periods; and devoted some time to asking how you might summarize these measures.

One of the measures, for example, that we talk about, and which is used: When will the system go broke? And that actually has its own loadedness in the language. Because, for example, SMI can't go broke, by its nature.

But in any case, for many of those measures, what's going on in 75 years isn't as important. Once you decide that the system is really going to be under strain in the year, you know, 2015 or 2025, or whatever the number happens to be, you don't have to worry quite as much in a policy arena about what's happening well after that.

And I think there was some concern on the panel for how these are presented and how they're used. And it's very hard to summarize complicated issues for a long period of time in sort of one number, without the sort of thoughtful comments that you made.

And we hope that people read the comments. We were worried at some level that people would view the report as alarmist, because when you change the growth number from 0 percent above GDP to 1 percent above GDP, things look worse. But we didn't view that necessarily as a cause for alarm, per se, if you thought about that for what that meant, and when that would happen.

PARTICIPANT [In Audience]: Can I ask a question?

DR. CHERNEW: Certainly.

PARTICIPANT [In Audience]: What you did was increase the size of numbers [inaudible]. And doesn't that cause you some concern?

DR. CHERNEW: Actually, we increased all of the numbers because--

PARTICIPANT [In Audience]: No, but concerning the 2025.

DR. CHERNEW: No--Well, I understand. Although, actually, some of the earlier numbers were probably increased as well. And our opinion was, we would rather come up with a number that we thought was more accurate, and worry about how that was interpreted, than intentionally down-weight a number and have a number that was inaccurate in terms of a point estimate. And essentially, that's how we resolved that issue.

We didn't think that inherently in the report those numbers had to be over-weighted.

MR. GINGRICH: Can I ask a question? We're not going to get into the details of this report.

DR. CHERNEW: Well, it does look like--

MR. GINGRICH: There are a number of people who probably wonder about the weights and the points.

But if I can take your point for a second, are you suggesting to me, or to us, that the group looked at 75 years, picked an accurate end number, and then worked out a percentage that would get you to that number? Or that you figured out a percentage you thought was accurate, and then accepted whatever number that percentage drew?

Because I'm confused as to which thing you were changing. And it sounds to me like you intuited a number somewhere, and I'm curious what you were intuiting.

DR. CHERNEW: We essentially, very roughly--We looked at historical numbers. We looked at a bunch of other factors. And we took half of the historical growth rate in medical care costs and said, "That's along--" So we took a number that we thought was low. In fact, Mark McClellan actually in his recent work has a number that's somewhat higher than ours.

We projected that number out for 75 years, to come up with the share of GDP devoted to health care. It turns out to be 38 percent, which makes most people gasp--So now is the time to gasp.

So 38 percent of GDP devoted to health care. Then we asked the question, "Is that sustainable? What does that mean for the economy? What types of mechanisms adjust?" Now, of course, this has to be done under current law, so it's not clear all those mechanisms could come into play. But essentially, that was the exercise.

And we decided, based on productivity estimates, that it was sustainable to have that level of cost growth. And we didn't worry about how policymakers would interpret that or weight those different numbers; although there is some discussion of presentation.

MR. GINGRICH: Okay. Yes, ma'am. Just a second. If you will slow down long enough, they will bring you a microphone, which will help us in taping this.

PARTICIPANT [In Audience]: Okay. Thank you. I'd like to continue this discussion, but in a different direction.

DR. CHERNEW: Okay.

PARTICIPANT [In Audience]: Because I'm familiar with what you're saying about the actuaries; that they take this work very seriously, and they're very highly respected. And I've read some of the discussion. And I know this is a very difficult issue. But I'd like to bring it back to Newt's point about, what can we learn, in terms of what more information we need, and then maybe at some future point, figure out how to get it.

For example, I understand, I think, from what you've said and from other discussions I have had that there is no assumption by the actuaries in terms of what the shape of the future system will be; i.e., the impending impact of diagnostics, the trend certainly toward less invasive surgery, away from surgical suites, very labor-intensive health care. Is that right?

DR. CHERNEW: Yes, that's right. And I can comment on that at greater length, but, yes. Essentially, the view was, even for the intermediate run, that it's very hard to predict what specific technologies would be used. It's very hard, even if you know what those specific technologies are--how they will be applied.

PARTICIPANT [In Audience]: Right.

DR. CHERNEW: If you look at a lot of the forecasting of, you know, PTCA [ph] versus cavage [ph], you might get in it.

PARTICIPANT [In Audience]: Sure, or versus cavage with stints versus [inaudible].

DR. CHERNEW: Versus cavage with stints versus--now doing it for kidney blockages.

PARTICIPANT [In Audience]: Right. Right.

DR. CHERNEW: So the answer to the question is, yes.

PARTICIPANT [In Audience]: Yes. And then, so there's that. Then, in terms of what you're talking about in terms of historical trends with very labor-intensive health care, there's no factoring--I mean, you're taking the last 25 years and trending forward in terms of workers engaged in health care, I think you said.

DR. CHERNEW: Well, we were doing that mostly to figure out the sustainability of the number that was picked based on sort of historical analysis and some of these studies, Mark's studies.

PARTICIPANT [In Audience]: Yes. Right.

DR. CHERNEW: David Cutler, who is on Mark's study, was also on the panel. And David has a study that argues that roughly 50 percent of cost growth is due to technology, which is roughly one of the numbers that Mark Pauly put up.

And we looked at the literature on the role of changing financing system on cost growth, which is a personal interest of mine. And we essentially came to the conclusion that it is not fruitful to try and go technology by technology and application by application to figure out what's going to happen once you get out beyond a very short window of time; and that it is more fruitful to look historically at what happened.

It is certainly true, and no one on the panel would argue that it is conceivable that there will be technological breakthroughs that will break the historical relationship between health care cost growth and technology.

The simple question is: Did we view those as the types of assumptions that were most plausible? And in part, the problem becomes that even when you solve them--The polio example I think is a wonderful example. Because although that reduced a whole class of expenditures for people with polio, what that did for health care spending overall is a somewhat separate question.

PARTICIPANT [In Audience]: Yes. Well, if I can just ask one more quick question, and then just a comment. There has been research done by Kenneth Manton [ph]. You may be familiar with that.

DR. CHERNEW: Right. Right. From Wisconsin?

PARTICIPANT [In Audience]: And that is that the projected rate of disability among people over age 65 is dramatically less than what had been projected in 1982.

DR. CHERNEW: Right.

PARTICIPANT [In Audience]: And I think you can see what this simple non-economist's mind thinks in terms of maybe projecting it forward, in terms of expenses in Medicare downstream.

DR. CHERNEW: Right.

PARTICIPANT [In Audience]: And then, second, his work shows that the cost of the last two years in life, at age 67, if you die at age 67, is seven times more than the cost of care in the last two years of life when you die at age 90.

DR. CHERNEW: Right.

PARTICIPANT [In Audience]: So to your point in terms of, I think, if I understood it right, assumptions in terms of bringing more people into the health care system, if you're looking at a health care system going forward, where we are in fact providing more care to people on Medicare, that would seem to have some impact down the road, too, in costs.

DR. CHERNEW: Right. Again, I'll try to be brief. I am familiar with Ken's study. In fact, there's a study at Rand looking explicitly at this issue of health status and spending.

But we did recommend in the panel that there be differential cost numbers given to survivors and decedents in every year of life, and so the mortality table--Right now, what happens is, there's a cost at a given age. It doesn't reflect the number of people at that age that pass away. And those people, as you mentioned, from Manton's work, are more expensive.

And so there is a recommendation to have a distinction between survivors and decedents at every year, so mortality health improvements are incorporated. And then we further had a research recommendation to expand the understanding of the relationship between health status and spending so future models could incorporate that. And there's actually an explicit effort going on in modeling some of that at the Rand Corporation.

And I think Manton's work is exactly on point for understanding how medical technology changes health status, and how that health status affects expenditures, and how technology influences all of that.

PARTICIPANT [In Audience]: Okay.

DR. CHERNEW: So there was a big call for that in the report research.

PARTICIPANT [In Audience]: Good.

DR. CHERNEW: And I agree with you.

PARTICIPANT [In Audience]: Well, and I thank you. And the reason--

PANELIST: Can I make a comment on this?

PARTICIPANT [In Audience]: Well, if I could just say something, and then I'll sit down. The reason I'm drawing this out is because in this town I think the only people that really are respected and looked to for some of this data indeed are the HCFA actuaries.

And as well intended as I know the work is, it tends to be very influential in terms of the general public, in expectations for how much money should we make available for future Medicare needs. And in fact, that is exactly how that study was spun.

DR. CHERNEW: Right.

PARTICIPANT [In Audience]: There were some news stories to the effect that this means that the prescription drug benefit, for example, may not be able to be as generous as has been expected.

So any further work any of us can do to support additional data I think would be very terrific.

MR. GINGRICH: Okay. Mark, do you want to comment?

DR. PAULY: Yes. I was actually on the predecessor committee, a technical advisory panel in 1991, which I guess bequeathed to you this assumption about health care growing at the rate of real GNP.

Actually, that panel mostly worried about--It was Medicare and Social Security. We mostly worried about the assumptions about the rate of growth of real covered wages per worker. And I think we concluded it was going to be real low. And then, of course, it turned out we were completely wrong.

But I wasn't particularly upset with the actuaries' assumption. And we actually have an expert on the panel here who could comment on this. Because my belief was that the decision on how much Medicare would spend at some time in the future is a political decision.

It doesn't really depend very strongly on what the technology is. It depends on how much representatives of the general public think the taxpayers are willing to spend. That's influenced, of course, by what you can get for the money. But ultimately, health care I think is not a product whose price is determined by the cost. It's a product whose cost is determined by whatever somebody is willing to pay.

MR. GINGRICH: Actually, as somebody who once practiced in that field, let me suggest to you that's not totally accurate. The Congressional Budget Office decision to score the Clinton plan as a tax increase in 1994 killed it. Because it was unsustainable in the popular media to go back home and explain that you're going to have this massive tax increase.

If the Congressional Budget Office had said, "No, this is simply a transfer of your insurance premium, and is therefore a fee, not a tax," it would have had a very significantly different impact on Capitol Hill.

And at the time, I can just tell you, there were a number of members who felt emboldened to oppose the plan almost overnight, and who were waiting for Reischauer to make the decision. And the second he said, "This is a tax increase," they felt very comfortable opposing it.

If you're not sure what to do--and most of the time in most of politics it's either real obvious, in which case there's no debate, or it's not real obvious, in which case you by definition don't know for sure what to do because it's not obvious--in that setting, if you have an authoritative figure who creates a framework within which you have to discuss something, it has an astonishing impact.

Having been in the room, having negotiated a lot of these bills, and having helped craft a lot of things, I cannot overstate, between the HCFA actuaries, the CBO scoring, the OMB scoring, in this city--maybe in the last month of an election it doesn't have any impact--but in this city, as the Legislative Branch and the Executive Branch wrestle with each other, without regard to party, setting the terms of the wrestling is a really big deal.

And if you say, for example, "Medicare will go broke, and you actually may need to raise Medicare fees or ration care. How can you think about a new drug prescription?" you're in one world. If you say on the other hand, "We have a sea of money coming in, and an extra couple of trillion to spend. How can you deny drug benefits?" you're in a totally different environment.

It really is psychologically that big. And the political will then is expressed in a sense through the kaleidoscope created by the prism of that kind of seemingly artificial accounting.

Did you want to comment?

DR. CHERNEW: Well, I want to make a very quick comment to Mark's comment. One of the very complicated things in this process was to understand what's under current law, and what parts of benefits you think are parts of current law which is what it's supposed to reflect, and what parts aren't. And things like the prescription drug benefit were considered not part of current law. But the idea of raising taxes was considered something that would have existed under current law.

And the explicit assumption of the panel was that historically people have demanded access to these medical technologies. And we essentially assumed that the financing and the taxing would respond to that and, if it wouldn't respond to that, we would be more apt to consider the not responding to that demand as a change, as opposed to the other way around. But that's a very complicated--

DR. PAULY: Well, I understand the process. I'm just asking about the realism.

DR. CHERNEW: All right. I see.

PARTICIPANT [In Audience]: I think I'm a futurist to some degree. But when you build a castle in the air that's 75 years out, somebody has to come and collect the rent. And so let's just cut the time down a bit, and look at what we just see.

The history of medicine is not a linear extrapolation. It's the history of revolutions and discoveries: polio vaccine, all the cardiac things you can think of, and so forth. We've just experienced one called the "Human Genome Project." That was done mostly by non-physicians. All the people--computer scientists, mathematicians, chemists, and so forth, making machines and so forth--got us this beginning of a revolution.

That will lead us to--inevitably, almost no way out--a change in all of medicine--almost all of medicine--from curative treatment to preventive treatment. How are you factoring in over the next even decade, or 15 or 20 years, that major shift in paradigm, in terms of measuring what you call "productivity"? And that goes for everybody on the panel.

MR. GINGRICH: No, but I think your case is particularly relevant. You're sort of like describing patterns of candle use in the year before Edison invented the electric light.

DR. CHERNEW: I think the answer, loosely, at least for the panel, was: "We agree completely, and we're not factoring it in at all."

In other words, I don't think that that's necessarily--You know, even looking at the genome example, which was discussed at some length, people were very uncertain about what the ramifications would be. Because it's not just that you have the genomes. It's how it's used. It's when it's used. Who's going to apply these? There were a lot of people that thought that would increase expenditures dramatically, and other people that thought it would decrease expenditures phenomenally.

And so it's not that we disagreed; it's we were trying to pick a number--And I will agree that looking backwards is no way to drive a car. And there was some aspect of that, admittedly.

And what we've done--Although I will say historically, despite the fact that the nature of the specific technologies has been dramatically different, the general gestalt has been not that different.

PARTICIPANT [In Audience]: But Mike, don't you think--I'm assuming; well, I hope you think--that new technology is basically--I mean, it's not determined by God, of course. Discoveries are made, but it's largely endogenous. If there's the money to pay for it, somebody will discover some good way to use that money.

And that's what I think makes it very hard to forecast. Because in addition to looking out 75 years, you know--

DR. CHERNEW: Right.

PARTICIPANT [In Audience]: All economists are told who are thinking of forecasting is, either pick a year or pick a number, but never say the same--

[Laughter]

DR. CHERNEW: Right. I'm going to let someone get another question.

PARTICIPANT [In Audience]: But I think the endogeneity of technology is a serious problem here.

DR. CHERNEW: Right. And I will save that for the next three questions, but, yes.

MR. GINGRICH: But let me just again make a passing comment on this, which is you may not understand what these numbers look like. There have to be envelopes of cultural adaptation. That is, when you have the horse, you paid "X" percent of GDP for travel; when you got to the automobile and the airplane, it probably went up fairly rapidly from 1905 to about 1930. My guess is, it has not increased nearly as rapidly since about 1930.

And so to some extent, it is true, all exponential numbers get very big over time. But that's all you've told us. I mean, unless you honestly believe--Although the other side is, if you have such dramatic increase in wealth that I want to pay the "X"-dollar for the, you know, cosmetic treatment, so I'm at the optimum level of feeling good the day before I die at 170, maybe you get to 30 percent.

As a historian, I would be startled if you couldn't go back to early stages of whole zones and, in theory, extrapolate similar patterns. You know, the rise in television cost in the 1950s, for example.

Back there in the back.

MR. DILLON [In Audience]: Ken Dillon [ph], Spectrum Bioscience, for any of the panelists. I'd like to introduce a concept that we haven't discussed yet, and that is appropriate technology.

Think over the past five years how many times you've run across the concept of high technology or advanced technology. And then think how many times you've heard of or thought of the concept of appropriate technology. Pretty dramatic difference, I think.

We generally think of appropriate technology in terms of something like a solar water pump in the Third World, or something like that. But if you were to apply it to medical technology, perhaps you could come up with a dramatic breakthrough.

But there is a problem. If you come up with your putative dramatic breakthrough, and you go to NIH and you ask the science bureaucrats--or better yet, the peer reviewers who are going to perhaps write the approvals for you getting a grant--you'll get bounced out of there.

Because they are dealing with advanced science. They don't appreciate the value of appropriate technology. I wonder if you have any ideas about what might be done to change that?

PANELIST: If we're lucky, the technologies are competing amongst themselves, and so are their proponents. And I would think that on the whole, most new technology would be improved technology, be recent technology, because that's how you gain some kind of advantage in this marketplace.

And as far as NIH is concerned, I mean, their job is to explore new things, rather than to figure out better ways to apply old things. That's someone else's job.

PANELIST: Well, I guess I really don't know very much about this. But I did read a McClellan paper which looked at the cost effectiveness of changes in treatments for heart attacks. And the biggest bang for buck was administering aspirin. That's clearly appropriate technology, by anybody's standard. And I believe there were a number of studies that were financed by NIH to get that result.

So I'm not sure that we're always only investing in high-tech and never looking for appropriate technology.

PARTICIPANT [In Audience]: Let me raise a question that we discussed quite a bit at lunch and that hasn't come up yet. We've been talking about certain kinds of technologies, but not very much about the fact that in the whole health care area there's been almost no application of communications technology as it is known in almost any other area of the economy.

Doctors still play telephone tag. They think electronic communication is by fax. They can't talk to each other. They can't talk to specialists. You know, the whole story--Everybody that I've ever talked to has a story about this.

I wonder if any of the panel would like to comment about this, what I think is an obvious failure of communications technology which is out there in almost every other industry--trucking, you know, you name it--why it has not had any significant effect so far in health care.

PANELIST: Well, I offered some thoughts on that about having to do partly with the sociology of medicine and the selection of practitioners in the field--And I think I agree with myself.

[Laughter

PANELIST: But I think, also, you--Well, this is almost a panglossian type of statement. But there's nothing I can see that would prevent you or me or anybody here, except investors, from creating an HMO that would use this new technology.

And at least in Philadelphia we have doctors coming out of the woodwork, so we could certainly sign some up if we offered them part of the action and promised not to harass them too much, if they would just adopt this new technology.

And it's kind of the market test argument I was making. And here I'm going to express ignorance, rather than knowledge. If there is a glitch here, I can't see what it is. And I actually once wrote a paper whose title, "What's Your Problem?"

There is certainly a lot of good intuition that would say just what you said, that there must be some ways to use this technology that already exist elsewhere in the economy and the health care sector, and somehow it isn't happening. But I have not been able to find sort of the bacterium that is somehow preventing that from going on in the health care sector. I'd be interested if anybody has identified sort of the negative active ingredient here which somehow seems to inhibit that.

MR. GINGRICH: Let me comment for a second. I think that it is--And I gave a speech here a while back that I think we passed out about sacred cows in health care. And I think health care is fascinating in part because it's the most complex ecosystem in our economy.

There's no other area that has as many different centers of activity. There's no other area that is both as decentralized and as politicized simultaneously. So that every state has all sorts of lobbies; every part of the guild has its own sense of self-importance.

Medical schools teach in a way which optimizes the sense of arrogant isolation; so that by the time you leave medical school you wonder why all these other mere mortals should be out there. You may remember the joke about the person who arrives in heaven and is standing in the cafeteria line waiting. Someone comes rushing through wearing a surgical gown, goes to the front of the line, grabs food and leaves. And he says, "Who was that?" And they said, "Well, God likes to pretend he's a doctor."

[Laughter]

MR. GINGRICH: The whole pattern here starts with the fact that we had an organic growth from around 1910 of a system of medicine which is doctor-centered. It's not a health care system; it's a doctor-centered system. And it's not an illegitimate system; and this is not about bad people. It's about systematic habit over a long period of time that has been encrusted with politics, which Adam Smith warned about if you read The Wealth of Nations. He talked about the degree to which every guild resists modernizing change. Why would you think medical guilds are different than other guilds?

Okay, second: Knowledge is power. And it's danger. I go back to the point that Mark Pauly is making. My insurance company wants me as a provider to give them all the information, so they can control me? Fat chance.

Third: All of your relationships are episodic. Somebody commented to me earlier today that they recently had the experience of going to a medical center for their daughter, and had to fill out the same information three times in 30 minutes.

Now, I'm just suggesting to you, this is a pattern which in any other aspect of American life would seem ludicrous. But the depths of the opposition to change are enormously deep, and morally self-righteous. These are the people who save your life. Are you going to look these people in the eye, as they're saving your life, and get them mad at you?

And so, I've been using the example of Firestone Tires, which has ten people a year who die of Firestone tires; and the Institute of Medicine report, which is between 40,000 and 92,000 deaths a year from medical error. Assume it's grotesquely exaggerated. Let's say it's only 10,000 deaths a year. Okay, that would be a thousand people dying of medical error--in hospitals only--per every person who dies of Firestone tires. But Firestone Tires is pretty easy to beat up. We have a culture that says business are greedy and evil; and we have a culture that says doctors are basically good.

So let me go back to your point. There are places doing it, and I've been in three places: Hippocrates, and I-Scribe on the West Coast, and Parkstone in Miami. All three provide hand-held devices for doctors; give them formularies, for example. Because the way you handle having multiple insurance companies for your patients is, you have a computer that does it; you don't try to do it. You don't remember all the stuff.

Hippocrates does it all by information on the Internet, and downloads it. They've got over 100,000 doctors who now have downloaded the data and use it regularly. I-Scribe and Parkstone actually have a device which orders electronically, and you can actually send an electronic prescription, etcetera. They are more expensive and more sophisticated. They're at about 3,000 doctors, I think, that are now using the Parkstone version. But they're starting.

But here's the challenge, one of the reasons that I got involved in trying to look at this. You cannot expect the medical system to voluntarily evolve, because the core requirements are standardization.

What do I mean by that? If every person in this room had an electronic medical record--which is what the technology would permit today--so that if you were in a different state on a trip and got sick, they could make sure not to give you the medicine that would kill you based on your medical record, to get to that, if you don't have the Federal Government require it, it won't happen.

It is not conceivable. Because you have liability laws; any single hospital can decide they want their own system. And that's why I suggest to you that sooner or later at some point we have to bite several very tough bullets, or the system will continue to be an essentially paper-dominated system, both on the transaction side, which is a total absurdity, compared to ATMs and gas pumps, and on the information-keeping side, which is actually dangerous to our health.

But I don't see any likelihood of it changing voluntarily, or being plausible.

PARTICIPANT [In Audience]: Can I make one comment, just to move this along?

MR. GINGRICH: Sure.

PARTICIPANT [In Audience]: I know of at least one study that looked at the extent of hospitals that computerized their records with some outcomes. And it has a positive effect, which is kind of what you'd expect. But the point is, it's been documented in at least one study.

MR. GINGRICH: Good.

PARTICIPANT [In Audience]: [Statement Inaudible.]

PANELIST: Oh, well, but consultants always say that.

PARTICIPANT [In Audience]: [Statement Inaudible.]

MR. GINGRICH: Yes, sir.

PARTICIPANT [In Audience]: All of the discussion has been around the cost of medical technology, with some allusion to the outcomes, the results. I would submit that a lot of our problem is because we have very good accounting for the cost, and lousy accounting for the benefit to patients.

Technology of, you know, administering for bacteria in one's stomach to cure ulcers: very cheap technology, has a lot of benefit, but there's no benefit to the patient that is collectively accounted and made public. And that, I think, is an issue that deals with the structure that you talk about.

MR. GINGRICH: Also, if I might, you also inject [inaudible], which is that we don't have a system that tells us, "This is the data we ought to be gathering."

PARTICIPANT [In Audience]: Well, it seems to me that HCFA could begin to do that. HCFA has full responsibility for end-stage renal disease. Why doesn't it do that, you know, as a matter of case?

PANELIST: Actually, it does do it pretty well for end-stage renal disease, but not a lot of other things.

PARTICIPANT [In Audience]: Well, but is it a public thing? Do individuals--

PANELIST: Well, it depends on what kind of--Yes, actually, there is something called the "United States Renal Data System"; although if you got their report you would want to put it there with the phone book. It's very thick.

But I mean, the problem I see is what you can get from that are measures of sort of physical outcome; but the valuation question, which is what you need to be able to compare the costs and the benefits, that's one where there's very little information.

PARTICIPANT [In Audience]: Well, cataracts, you can measure the benefit in quality of life over a period of time, for improved vision. And that's a documentable thing.

PANELIST: Right. But how much is it worth?

PARTICIPANT [In Audience]: You've got to assign something someplace. Maybe it's a percentage of life. But you have to have it. Without it, you don't get anywhere.

PANELIST: Well, I agree with that.

MR. GINGRICH: Okay. Yes, sir. Wait for them to get up to you with the microphone, if you don't mind.

PARTICIPANT [In Audience]: Some of the concerns I've heard about communications technology, use of appropriate technology like aspirin, prevention, and HCFA--I do have one little piece of good news, actually. HCFA is in fact starting later this year a demonstration project in an area which is extremely expensive, utilizing all of these things that we've just talked about, and it's for heart disease.

We're doing it, so I know a lot about it. But this leads to, there's an interesting thing in Mark McClellan's paper which I don't know if he covers. But cardiology or heart disease is not just a disease; it's a practice problem. And part of the problem is what in bad parlance is "over-utilization."

And over-utilization can actually be demonstrated. The president of the American College of Cardiology introduced Jack Winberg [ph]. I don't know if you know him. He keeps the Dartmouth Medical Atlas. Turns out, shockingly, that the second-best predictor of whether you're going to have an invasive procedure is your zip code.

And what's intriguing about that is, depending on what zip code you live in--and this is amazing--depending on the zip code you live in, whether or not you're going to have a bypass, or a coronary angiography, or an angioplasty, varies to 200, 400, and 1,100 percent.

Now, I don't know if Mark covers any of that in there, but the fact that that exists indicates that there's a problem with the way medicine is practiced on a geographic basis. And you don't have to be a bad doctor to practice medicine in a vigorous way. You might have liability concerns. You might have peer pressures. And of course, you have what you learn in medical school.

But nevertheless, there is some hope. There's some hope on the way.

PANELIST: Mark's paper does make some brief reference to the fact that these studies do not look at basically the allocation of the new technologies, where they're used. They looked at sort of--at averages.

But I don't think we should lose the larger point, which is the steady advances of these technologies in general. I would imagine, at virtually every zip code, the chances of getting invasive heart treatments right now are much harder than they were ten or 20 years ago, probably with a lot of benefits.

PARTICIPANT [In Audience]: Well, yes. And if doctors are so bad and they don't know what they're doing, then what did account for these substantial drops in death rates from heart disease and stroke?

PANELIST: Right.

DR. CHERNEW: But that claim is exactly the distinction between marginal and average productivity. At the margin in these very high-use places, you might be getting something different than on average, under and over.

PARTICIPANT [In Audience]: It could be over or under.

MR. GINGRICH: Well, but let me ask just for a second. I don't think it's a question of individual doctors being bad. I think it's a question of cultural patterns.

[Tape Change]

PANELIST: We've got a mythology that dominates our thinking about medical care: that the invention of all the great vaccines and treatments for contagious medicine was responsible for the decline in the death rate from contagious medicine. It isn't true. It is responsible for a very, very small proportion of that decline.

And so what you get in looking at medical treatment is that there's other things going on out there that we don't understand very well. That just complicates this whole issue of what data we're going to get.

There are other things that affect the outcomes that are not--either we don't understand them and they're sort of outside the medical treatment, or I don't know what--that tend to have a very strong effect. And we'd like to understand those things as well, to understand the real effect of medical intervention. This makes the whole problem more complicated.

PANELIST: I think even with Medicare, we have the structure in place to solve this problem. You create the Dr. Ornish prudent health care Medicare HMO.

PANELIST: Right.

PANELIST: And it uses the money it saves to provide prescription drug coverage. And I need to ask myself, "In six years, is that the HMO I want to join?" Because the answer might be "Yes," or it might be "No." It depends on how much different plans--

MR. GINGRICH: But you start with your first point. We spent a year and a half--Ornish spent; I wasn't directly involved. Ornish spent over a year and a half to get HCFA to agree that they would pay for 1,800 tests nationally. And their first question was, "Can you really prove that diet and exercise are safe for senior citizens?"

Now, I'm just suggesting to you, to get them to agree to it--I mean, I suspect Ornish tomorrow morning would be glad to set up an HMO, if they pay for it.

PANELIST: Well, but the HMO takes its AAPCC, and it can pay for it.

PANELIST: Right.

MR. GINGRICH: No. Not in terms of--Well, I mean, I think that you'd find real resistance at HCFA if you walked in and said, "This is what I'm going to do."

PANELIST: I think that's probably right.

MR. GINGRICH: Yes. Okay? One or two more questions. We don't want to keep you all too long. Way in the very back, the lady way over there.

MS. CROWDER [In Audience]: Hi, Joelie Crowder [ph], with the American Health Quality Association.

Actually, just a couple of comments: We represent the Medicare peer review organizations--which there's a lot of HCFA Medicare bashing going on. But it is a very small program, relatively speaking, compared to the budget, the entire Medicare budget. And it's a program that--The focus is health care quality improvement. And it's not very well known that they're research based. They have 23 quality indicators that they're out working with docs out in the communities. And just to point out that there is good work being done. There's not a lot of money, like I said, relatively speaking, focused on this program.

And the other issue is to talk about Coltry's [ph] book about the culture in medicine. And it's an issue that we're facing right now with the peer review organizations. We're struggling with making a business case for quality.

And this also goes back to your comment about appropriate care. Everything that they're pushing--the 23 quality indicators are evidence-based, research-based indicators that are appropriate care that patients should be receiving. But they're going out in the communities and meeting physicians in their own health systems and working with providers.

And they're meeting resistance to changing practices of care; to this, what they have defined as appropriate care. And it's the culture of health care which has said, "We need to make a business case. It needs to be a financial case. There need to be incentives in the system for meeting an improved quality of care"; which we've found very difficult and we're trying ourselves to work on to help make the business case.

MR. GINGRICH: Before you give up the microphone, is there a website for your particular activity?

MS. CROWDER [In Audience]: Actually, "WWW.AHQA.ORG <http://WWW.AHQA.ORG>" is our website, which links to all the peer review organizations; which there's one in every state and the territories. And then, "HCFA.GOV" is also another website. And if you search for "Quality" on their website, it has information about it.

And they sponsor a lot of different demonstration projects on different quality of care issues. And they're also a resource--The pros are a resource that have access to a lot of the data that researchers use, and could be a good resource for other folks who are interested and serve as resources in the state for people who are trying to get access to data, as well.

MR. GINGRICH: And one more thing. When you say you're getting a push back on adopting qualitative improvements in care unless it's financially profitable, does that push back come from insurance companies, or doctors, or hospitals?

MS. CROWDER [In Audience]: I would say the focus--We're working on developing a symposium for the spring. And we're work