March 6, 2008
Session 3: The Paradox of Plenty
Elliott S. Fisher, M.D. , M.P.H.
Center for Health Policy Research and Clinical Practice Dartmouth Medical School
CHAIRMAN PELLEGRINO : Thank you very much for returning so promptly, members of the Council. Dr. Fisher is on the phone. Do you hear me, Dr. Fisher ?
DR. FISHER : I hear you fine. Can you hear me?
CHAIRMAN PELLEGRINO : Yes, I can, very, very well. I think we'll proceed along the lines of the agreement we made. Dr. Fisher will give us a short presentation, describe his slides, and then we'll open it up to discussion. Thank you again once more for taking time out from your visit with your daughter to speak to us. We appreciate it very, very much. It's all yours.
DR. FISHER : Well, first it is a great honor to be here and to join you to talk about something that I care about deeply and to share some of what we've learned and some of the challenges we face.
If we go to the first slide, this is a graph of projected spending on Medicare and Medicaid prepared by the head of the Congressional Budget Office, Peter Orszag, and published in the New England Journal. Can you hear me okay when I speak this way or is there any problem? You folks are okay? Am I speaking clearly enough?
CHAIRMAN PELLEGRINO : Yes, you're speaking superbly.
DR. FISHER : Great. Thanks. I will ask occasionally, "Am I on the right slide?" I want to make sure we're looking at the same thing. So this slide shows projected spending on Medicare and Medicaid. The upper line reflects the current growth rate of spending in Medicare and Medicaid projected over the next 30 or 40 years. And what you see is, if current trends continue spending on Medicare and Medicaid alone will account for 20 percent of the gross domestic product, approximately the same share as is made up now by all of government spending. There are profound potential impacts on both the public and private sector from the increase in cost.
The same article by Peter Orszag highlighted the pretty remarkable variation that is observed in per capita Medicare spending across U. S. Regions, two-fold differences at least between the areas with the lowest spending regions, the pale regions, and some of the lower spending areas within the pale-colored regions, and much higher spending on some regions, reaching well over $10,000 per beneficiary in some of the higher spending regions, one of which is McAllen, Texas; another is Miami. I'll come back to those later.
This is largely based on work that Dr. Wennberg and I and others have done up here at Dartmouth . The focus of our work recently has been trying to understand the causes of these variations. Most recently we have now been able to look at the care provided by specific hospitals and their associated medical staffs, so when we published the Atlas of Chronic Disease two years ago, we looked at variations in the patterns of practice across academic medical centers, because it turns out that when patients are seriously ill they get almost all of their care from the institution where they first get their care.
In this particular case, in order to control for differences in illness across different populations and be sure that we're looking at similarly ill populations, we look at the care of patients with chronic illness who are in the last two years of their life.
In this graph each of the blue dots—are we all looking at Slide 4? Chairman Pellegrino ?
DR. ROWLEY : Three.
DR. FISHER : Slide 4 is up?
CHAIRMAN PELLEGRINO : Yes.
DR. ROWLEY : This is four.
DR. FISHER : I'm sorry, I might have forgot to say "next."
DR. ROWLEY : This is on medical care, which is what he—medical center, which is what he talked about before.
DR. FISHER : So what is shown on Slide 4—and I will try to use the numbers so the technicians should be able to see these on their screen.
DR. ROWLEY : [inaudible]
DR. FISHER : I'm sorry?
CHAIRMAN PELLEGRINO : If you're going to communicate with Dr. Fisher , use your microphone and put it on.
DR. FISHER : I missed that. Are we on Slide 4?
DR. ROWLEY : Yes.
CHAIRMAN PELLEGRINO : Yes, we are on Slide 4.
DR. FISHER : Okay. Thank you. So this is—each of the blue dots represents per-beneficiary spending during the last two years of life for beneficiaries who died between 2000 and 2003 and who had chronic illness at the time of their deaths during those last two years. The blue dots represent the hundred top academic medical centers in the United States according to AAMC, so you see almost three-fold differences in use of resources and Medicare spending across these hospitals. The red dots represent what are the U. S. News and World Report's top ten honor role academic medical centers during this period. The data for those hospitals are shown on the right.
So what you see is greater than two-fold differences in per beneficiary across what some might call the best academic medical centers in the world. So if we go to Slide 5—so next slide— Uwe Reinhardt looked at data similar to this and was motivated to ask the question, "How can the best medical care in the world cost twice as much as the best medical care in the world?"
That has been a wonderful motivation for our research. If we go to Slide 6, I'll set you up for what I plan to talk about over the next 40 minutes. First I want to talk about what we know, what the evidence shows so far, about the differences, the implications for health, the differences in spending across regions, then what we think we know and what this might mean for science and policy. I want to emphasize that the work I'm sharing with you has been broadly supported. We have had many wonderful funders, and I have numerous collaborators who should all be here presenting this with me.
So let's go the next slide. It should be Slide 7. The first major set of studies we did looked at patients who had heart attacks, colon cancer, hip fracture, or were participants in the Medicare current beneficiary survey, which is a random sample of the Medicare population over age 65. We took advantage of the differences in spending across regions because those essentially represent a natural experiment.
We were able to control carefully for illness differences across communities in ways that managed to persuade the reviewers. I won't bore you with the details. But essentially we convinced the reviewers that we would get as close as one could get, using what the economists would call instrumental variable analysis, to creating a randomized trial. We looked at the care of patients in the pale areas with those four conditions and essentially compared it to the care of patients in the areas of increasing intensity. We measured intensity after adjusting carefully for price and for illness differences across the region.
Let's go to the next slide, Slide 8. The first thing we looked at was the content of care. I want to spend a moment on clarifying three categories of services that we think are actually critical to understanding how health care works in thinking clearly about the unwarranted variations in practice that we see across U. S. regions.
First, effective care. These are services where we know all patients with a particular set of clinical indications should receive that specific treatment. There are no tradeoffs involved, so the example, if you're having a heart attack you want someone either to give you a clot-busting drug or you want them to quickly put in a stent or a little balloon to push the blockage that's in your coronary artery out of the way. In those cases the right rate across regions would be everybody having a hundred percent compliance with these evidenced-based guidelines.
The second category is what we refer to as a preference-sensitive care in a very normative sense. These are treatment choices that entail tradeoffs among the risks and benefits where patients' values and preferences, we believe, should determine the treatment choice. That does not happen to be the case currently in practice. Generally physicians tend to prescribe treatments that they think are the correct treatment, and we believe this is an important issue. It's not the major focus of my talk this afternoon. The example would be coronary bypass graph for stable angina where the benefits are improved angina for most patients, no general benefits in most cases in terms of life expectancy, but there is a trade off. That is, most patients will experience improvement in their angina, but about ten to twenty percent of patients will have some degree of cognitive impairment or harm as a consequence of the procedure. Therefore, this is a choice between whether you want to play tennis or whether you want to remember your daughter's name. That's a choice that we believe only patients can make.
The third category are services where utilization is strongly associated with the local supply of health care resources, like the frequency of physician visits, specialist consultations, use of the hospital or ICU at the site of care, and the other things that will go along with that.
Next slide. I think we should all be looking at Slide 9, "Reperfusion in 12 hours for AMI" with a little red dot there. Dr. Pellegrino , is that correct?
CHAIRMAN PELLEGRINO : Yes, it is correct.
DR. FISHER : Thanks. So at least you're still awake. Is anyone else awake?
CHAIRMAN PELLEGRINO : I have insomnia.
DR. FISHER : All right, well, thank you for my tolerance of my slight ribbing. I really wish I were there with you. It would be so much fun for me and for you, I think. However, let me explain this slide. It's a complicated slide, and I'm going to show you several others that look like it.
What this slide shows is the ratio at which this particular service, in this case an effective care reperfusion in 12 hours, is provided in the highest spending region, which is labeled "Quintile 5" (that's the fifth of the country that has the highest spending rates) compared to Quintile 1 (that is, the fifth of the country, the pale areas on that earlier map, that has the lowest spending rates).
49.8 percent of patients get reperfusion within 12 hours in the high-spending regions, and 56 percent of them get it in the low-spending regions. The dot is therefore off to the left, and this is usually where I ask everyone to raise their hands and answer the question how many people think they are doing better in the low-spending regions. And then, of course, I can't see, but Dr. Pellegrino will affirm that everybody is raising their hand.
This says care for this particular piece of evidence-based practice is better in the low-spending regions than the high-spending regions.
Let's go to the next slide, which should be Slide 10. What we see is that for four of the six measures of effective care, quality measures collected by the Medicare program from chart review in 200,000 patients, care was better in the low-spending regions, and it was no different for two of the six measures. In ambulatory care we see that for three of the four measures, care was better in the low-spending regions than the high-spending regions.
Does everybody understand these graphs? I see lots of people nodding. Good.
CHAIRMAN PELLEGRINO : Heads are bobbing.
DR. FISHER : Good. So let's go to the next slide, Slide 11, which many of you may have seen. Looking at the relationship between Medicare spending per capita on the horizontal axis and on the vertical axis the ranking of each state according to the Medicare program's comprehensive measures of quality that were developed by Steve Jenks when he was with the QIO program, we see what is a quite strong negative relationship on overall technical quality between increased spending and quality rankings. Spend more, technical quality, the provision of effective care, is worse.
Let's go to the next slide, so we should all be looking at Slide 12. This is for preference-sensitive care. This was one of the surprises that we found in the data. That is, if you live in a higher-spending region, getting about, this time, $3,000 more each year per Medicare beneficiary, you don't buy more major elective surgery after your heart attack, even any more bypass surgery.
Now, I'm going to take a little detour because some of you may have heard that Dartmouth looks at variations in practice across regions and has shown, as we do, four- to six- to ten-fold differences in rates of exactly the procedures that I'm showing you here across U.S. regions. So this may seem paradoxical, and I want to explain it for a second. So let's go to the next slide.
This should be Slide 13 and you should see five little groups of dots, which are either a Rorschach test if you tend to be a psychiatrist or a turnip if you're more of an organic farmer. So Quintile 1 to Quintile 5. Each of those dots represents the rate of coronary artery bypass surgery in the hospital referral regions that comprise each of those spending groups. So Quintile 1 would be the pale areas on the map that I showed you earlier, the lowest spending regions. Quintile 5 would be the highest spending regions.
We still see dramatic variation in rates of surgery within communities of differing spending levels. So marked variation in surgical rates for preference-sensitive, or elective surgery, as some might describe it, but no general overall increase in the rate of elective surgery in the higher-spending region.
Let's go to the next slide, Slide 14. That should be the old slide, and then we go to Slide 15, I hope—oh, no. This is a bill I probably got left in there. Can you hit the forward, and you should have still Slide 14 but with two groups of dots. Is that correct?
CHAIRMAN PELLEGRINO : Correct.
DR. FISHER : So what you see for major elective surgery, no increase in the provision of major elective surgery as we spend more across region. Very interesting and kind of fun finding. So all of you on the Council took plenty of multiple choice tests when you were in college or medical school. I set you up to say there are three categories of care, so you know where the money has to be being spent, in the higher spending region. Let's go to Slide 15.
What you see is 30 percent greater use of office visits in the red high-spending regions compared to the low-spending regions, twice as many inpatient visits per capita. This is among beneficiaries in the high-spending regions compared to the low-spending regions who had one of those three conditions and adjusting for their health status at the time of hospitalization. It's actually in the year after your AMI or, in the bottom graph, the last six months of life for all patients over the five years of follow-up.
So similar patients in high-spending and low-spending regions, 30 percent more office visits, two times as many inpatient visits, two and a half times as many inpatient specialist consultations, and four times as many visits to psychiatrists. They're not sure whether that means they're more depressed in the higher-spending regions or that they're more effectively treated for what depression they have. That study is still to be done.
They spend more time in the hospital, 30 percent more discharges, and each of those stays is slightly longer. So total inpatient days in the year after your heart attack are greater, and two times as many days in intensive care units. Again, identical patients, twice as much time in an ICU. And then, of course, if you're horizontal in a bed and more specialists are seeing you, you're going to get more minor tests and procedures, and these are just some selective procedures that we have shown. And, finally, in the highest-spending regions patients are treated much more aggressively at the end of life, much more likely to have a feeding tube, much more likely to undergo an attempted resuscitation.
Let's go to Slide 16. This summarizes what we know about the content of care, but we, of course, haven't addressed the problem of health outcomes. That was the major intent of the research, to understand whether all of this additions spending, which is mostly related to how frequently people are seen and how much time they spend in the hospital. Now we know that. Does that increase your length of life? And the answer, if we go to Slide 17, is no. In fact, slightly the opposite. That is, if you are living in one of the higher-spending regions, on average patients with heart attacks, colon cancer, and hip fracture are more likely to die in the five years following their initial event than those in lower-spending regions, 2 percent higher for the hip fracture patient and 5 percent higher for the heart attack and colon cancer patients.
We looked at functional status. There was no improvement in functional status and no slower rate of decline in the higher-spending regions.
Since then, if we go to Slide 18, we have now analyzed data from physician surveys with about 15,000 physicians practicing around the country. And what we see when we interview physicians is that those in the higher-spending regions compared to the lower-spending regions perceive that the overall quality of care, their ability to provide high-quality care, is worse. They are more likely to describe inadequate quality of communication among physicians. They are more likely to describe greater difficulty ensuring continuity of care. They are much—their overall assessment of their ability to provide high-quality care is lower in the high-spending regions compared to the low-spending regions. And my favorite finding of the whole ten years of work is that in these regions, which you will soon learn have 30 percent more hospital beds and 65 percent more specialists, they are more likely to perceive a shortage of hospital beds. That is, it's difficult to get patients into the beds. And in those regions which have 75 percent more medical specialists than interns on average, they say it's harder to get a new consultation with a medical specialist.
Part of this wonderful paradox of plenty, which has been such fun to try to unravel, when we look at patient-perceived quality we see lower satisfaction with hospital care and worse access to primary care.
Most recently we've begun to look at trends over time in health outcomes, and this is perhaps the most worrisome finding of the work that we have done. First let me show you what has happened across regions that differ in their growth rate. So let's look at—this is thanks to Google maps. This should be Slide 21, is that correct?
CHAIRMAN PELLEGRINO : No. Nineteen is what you have up there.
DR. FISHER: We'll let's charge forward. Nineteen should be patient-perceived quality. Let's go to Slide 20, which just sets you up to look at trends over time. Thank you, Dr. Pellegrino .
And now let's go to—this is hard to remember because I'm advancing my slides and I'm thinking that I'm—the reflex is to assume that you are. Thank you for correcting me. So let's go from Slide 20, which says I'm going to look at differences in trends over time across differing regions, right? We're going to look at what happens to those regions that grow the most quickly compared to those regions that grow in their health spending more slowly.
Let's go to the next slide and look at an example of what's happened in two cities. This should be Slide 21, and you should be looking at a map of southern Texas .
CHAIRMAN PELLEGRINO : Correct.
DR. FISHER : Okay. I'm going to compare what happens in Medicare spending between El Paso , Texas , and McAllen , Texas , both border towns, both in Texas , and let's look at what's happened.
This should be 1994. It's not shown on the slide. In 1994 spending, we see that utilization of hospitals, physician services, labs and testing on health care were quite similar in El Paso and McAllen . Understand the bar graph? You should be looking at Slide 22. These two communities were relatively similar early in the '90s.
The upper red bar on Slide 23 represents what's happened to the growth in spending in El Paso from—in McAllen in the upper bar, the red bar, versus the green, which is El Paso . Let's go to Slide 24 so we look at the most recent data. This is from 2003. What we see is per capita spending in McAllen on short-stay hospitals has increased dramatically, mostly due to increased utilization of those hospitals. Spending on physician services—
CHAIRMAN PELLEGRINO : We've got a disconnect on slides.
DR. FISHER : It's Slide 24?
CHAIRMAN PELLEGRINO : It's labeled 24.
DR. FISHER : And you do not have two bar graphs there?
CHAIRMAN PELLEGRINO : No.
DR. FISHER : What do you have?
CHAIRMAN PELLEGRINO : What we have is the "Paradox of Plenty," "Content/Quality of Care," et cetera, but no bar graph.
DR. FISHER : Okay. For the higher-spending region? Yeah. So let's go back for a moment to what I think is your Slide 23 now.
CHAIRMAN PELLEGRINO : We're on 23.
DR. FISHER : So that should have the two diverging lines, is that correct?
CHAIRMAN PELLEGRINO : No. It has a bar graph.
DR. FISHER : Oh, boy. All right, so we're going to be off numbers. So 23 there, is that—you have two bar graphs—
CHAIRMAN PELLEGRINO : It has bar graphs, "Short-Stay Hospitals," "Physicians"—
DR. FISHER : Right. And does McAllen look a lot higher than El Paso ?
CHAIRMAN PELLEGRINO : Yes, that's true.
DR. FISHER : Okey doke. So what you see is the growth in spending in McAllen is pretty wide, and these cities have diverged dramatically over this ten-year period. Okay?
Let's go to the next slide. I don't know what's happened to our numbers here, but it should say "The Paradox of Plenty," what the higher-spending regions get.
CHAIRMAN PELLEGRINO : That's correct. Twenty-four, "The Paradox of Plenty."
DR. FISHER : Great. So when we look at the bottom, "Trends over Time," we've looked at what's happened to survival in AMI in the rapid-growing communities compared to the slow-growing communities. We know overall there have been dramatic improvements in survival in AMI over this 15-year period. But what's very interesting is that the regions that grew the most rapidly have the slowest improvement in heart attack survival and so fell further behind the lower-spending, more conservative regions.
So let's go to the next slide, which should be a pop quiz if I'm really lucky.
CHAIRMAN PELLEGRINO : Correct.
DR. FISHER: So everyone has to do this and fill out their napkin. If all U.S. regions could adopt the practice patterns of the most conservative U.S. regions, the pale areas on the map, such as northern California or Rochester, which of the following statements would apply: U.S. health care spending would decline by over 30 percent; the projected deficit in the Medicare trust fund would be postponed by at least 25 years, and we could send 30 percent of the U.S. health care workforce to Africa and in theory improve the health of both continents; or, your last choice, all of the above?
I see everybody is raising their hands to say Number 4. That's great. So I believe this is in fact what would happen if we could do that, but we recognize that that is not going to be easy to do. But it should frame our thinking about the magnitude of waste within the U. S. health care system.
So my first major point, which I think will be your Slide 26—is that correct?
CHAIRMAN PELLEGRINO : Correct.
DR. FISHER : It should be "What I Know." This is a summary of the first section of the talk; that is, higher spending is largely due to overuse of supply-sensitive services like hospital and ICU stays, physician visits, and, at the margin, greater use of these services is worse.
So let's try to figure out what's going on. I'm going to—some of this is more hypothetical, some of it is grounded in some of the research we've done, but we're trying to tease out what are the causes in the differences in practice across regions and what might explain the worse quality we see in the higher-spending regions.
So let's go to the next slide. It should be a slide that says, "General Attributes of U.S. Health Care: the assumption that more is better," and I think this is a general attribute of U.S. health care which pervades all of U.S. health care at the current moment. This particular book by David Cutler , who has become a close friend as we've started to work together to try to think about the opportunities for reducing the amount of waste in health care. But Prof. Cutler has argued strongly that we do need to increase spending, that the increased spending we have made has resulted in dramatic improvements in life expectancy, and those improvements have been worth it.
I think this notion that we must spend more and must continue to spend more is a general attribute of U.S. health care. We'll come back to this again.
A second attribute well-recognized by all of us is the current information we have about current treatments, whether it's Vioxx. This is an advertisement that ran the very day that Merck withdrew the drug from the market. We have inadequate information on the risks and benefits of the treatments that we offer to our patients.
Third issue, going to Slide 30, I think there is a recognized and growing tension between science and professionalism on the one hand and the market forces that are driving U.S. health care.
Is it Slide 30 you have now? It's Slide 31? You're looking at "We do Botox"—somewhere there?
CHAIRMAN PELLEGRINO : Looking at 30 right now.
DR. FISHER : Let's go to the next one, which I think should be "Growing tension"?
CHAIRMAN PELLEGRINO : Twenty-nine. What's going on?
DR. FISHER : I think the challenge is somewhere your slide version and mine got separated, and I have—what slide number is this one?
CHAIRMAN PELLEGRINO : Twenty-nine now, entitled "What might be going on."
DR. FISHER : I have to mentally subtract one from my copy of the presentation.
My colleague Robin Larson looked at advertising by academic medical centers, and I only chose two to advertise here to emphasize the challenge of commercialism that confronts academic medicine at the current moment. University of Washington I chose because I trained there. The advertisement for Botox is quite startling. And then I chose the ad that they found to look at from Johns Hopkins about uterine fibroids because it was the most puzzling of the advertisements that we came across. None of us thought we would feel better knowing more about uterine fibroids.
Let's go to the next slide. I think there is a growing recognition led by many, including David Blumenthal at the Massachusetts General Hospital , of the conflicts of interest in academic medicine and elsewhere in the current environment. But these factors pervade all of the United States and do not give us much insight into the regional differences in spending. So we've been trying hard to sort those out.
Let's go to the next slide. It should be Slide 31: "Regional Differences in Practice and Spending. What are the underlying causes?" We have looked quite carefully at the role of patient preferences, how do patient preferences differ across U. S. regions, and they don't differ much. There is a light preference for specialist care in high-spending regions. There's no difference in preferences requesting tests if the doc says you don't need it, and there's no difference in preferences for aggressive end-of-life care, where we see the greatest differences in practice across U.S. regions.
We've looked for malpractice environment, and it does explain a little bit of a difference in spending across regions but no more than 10 percent of the two-fold differences we see. Where we think the problem lies is this issue of capacity and payment systems. Let's go to the next slide.
If we look at what should be Slide 33—you should have Slide 32, "Capacity and Payment" at the top, what are the regional differences, what are the underlying causes? If you push the advance, I think a little map should pop up, if I'm correct.
CHAIRMAN PELLEGRINO : That's correct.
DR. FISHER : Hallelujah. This is a scary process, Dr. Pellegrino .
CHAIRMAN PELLEGRINO : Scary for me, too.
DR. FISHER : You should have a map there reminding you. Pale areas, low spending; red areas, high spending. Let's push the advance one more time, and you should see a bar graph up here; is that correct?
CHAIRMAN PELLEGRINO : Correct.
DR. FISHER : Wonderful. And what you see is that we have the red bar represents the high-spending regions where they have, as I mentioned earlier, more hospital beds and—32 percent more hospital beds on a per capita basis and 65 percent more medical specialists.
Let's go, I think, to the next slide, and here you see a scatter graph, I hope—
CHAIRMAN PELLEGRINO : Correct.
DR. FISHER : — cardiologists on the horizontal axis, cardiologists per capita, and the frequency of cardiology visits on the vertical axis. So a strong positive relationship. The more cardiologists you have per capita, the more likely you are to see a cardiologist.
Let's go to the next slide. This should be a graph that has some grayed-out stuff, and we're talking about the capacity-and-payment system because capacity is strongly correlated but interestingly explains less than 50 percent of the differences in spending across communities. Because of the second factor, we do have a payment system that keeps us all fully occupied.
Let's go to the next slide. Oh, gosh. I apologize for the misspelling if "evidence" is missing an "e" on your slide. The question is, what's the role of physician judgment in driving the differences in spending or in practice across regions. What we did is , we used clinical vignette to sort this out. And we found that in clinical settings were there is strong evidence—the evidence is black and white, the trials are strong—there is no difference in answers across regions of different spending levels. People are likely to day "do it" at about equal rates in high-spending and low-spending regions.
But if go to the next slide, when we look at decisions where what we call the gray areas—discretionary decisions—where there is no evidence to guide practice, physicians in the high-spending regions make very different decisions. For instance, in this particular vignette the question was, "In general, how often do you schedule routine follow-up visits for your patients with well-controlled hypertension?" Physicians who practice in low-spending regions, only 22 percent of them would say, "I would need to see them in three months or less," whereas almost half of those in the high-spending regions said, "I should see these patients in three months or less."
We were clued into this question by running focus groups of physicians in Miami , the highest-spending region in the country, and Oregon , one of the lowest. And in those focus groups the physicians in Oregon said, oh, we would never see our patients—maybe once a year, maybe every six months, which reflects the data that we received from the national survey when we did it.
When we asked the physicians in Miami that question, they almost all said, "I would see my patient with well-controlled hypertension every month." So where physicians—and we found this to be generalizable to other clinical vignettes where judgment was a key element of the decision.
So there's no difference in intervention rates when decisions have strong evidence, but physicians are much more likely to intervene in the gray areas. In the high-spending regions they're more likely to refer to a specialist, more likely to see a patient more frequently, and more likely to admit an identical patient—that is, a patient in this clinical vignette—to a hospital.
That should have been, I think, your Slide 38. I'm sorry if I failed to mention advance it, so this should be the summary slide, Slide 38.
CHAIRMAN PELLEGRINO : Thirty-eight.
DR. FISHER : Is that correct, 38?
CHAIRMAN PELLEGRINO : Yes. It's entitled "Regional Differences in Practice and Spending."
DR. FISHER : Perfect. If you go to the next slide, you should see one more line added at the bottom about trainees, is that correct?
CHAIRMAN PELLEGRINO : No.
DR. ROWLEY : Yes.
CHAIRMAN PELLEGRINO : "Trainees"?
DR. FISHER : "Trainees" at the bottom.
CHAIRMAN PELLEGRINO : Oh, yeah. Right you are. Correct.
DR. FISHER : I heard a yes and a no. I'm—
Male Voice: You're there. You're fine.
DR. FISHER : Thanks. So our most recent study, which has been published as an abstract and which is now under review because the abstract findings were published, was quite interesting. We looked at trainees in internal medicine. We developed a score that assessed their ability to practice conservatively when doing the right thing meant watching the patient or withdrawing therapy or doing the least invasive treatment when the options were—this is on the national exam that all internists take. And we found that trainees in higher-intensity environment were more likely to make mistakes of overuse. They were less likely to do the appropriate conservative treatment.
So what's going on? Let's go to the next slide, which is my summary slide of our hypothesis about what is driving the differences in practice across regions. What I think is going on is that the clinical evidence—so does this say "Likely Diagnosis," your slide?
Male Voice: Yes.
DR. FISHER : Hallelujah. So I think the likely diagnosis is that local capacity in clinical culture drive practice and spending. Clinical evidence, principles of professionalism that we were trained with are critically important, but they are in fact a limited influence on clinical decision making. Physicians practice within a local organizational context and policy environment that profoundly influence our decision making. Payment system ensures that any existing capacity and any new capacity we bring in, whether we recruit more cardiologists or more orthopedic surgeons or whether we build more hospital beds, as in McAllen , Texas , the payment system ensures that that is fully utilized. And then that capacity helps determine local culture.
What we believe is happening is that because—for a primary care physician, for example, managing a patient with depression or managing a patient with back pain, in the current payment environment where you cannot afford to see a patient for more than 10 or 15 minutes, if there is an orthopedic surgeon for the patient's back pain, you will be much more likely to refer that patient. So whatever capacity you have in place is taken advantage of. Moreover, if there are lots of specialists in your community, we are likely to learn a specialist-oriented pattern of practice. But the consequence is that where there are reasonable—that is, these gray area clinical decisions, reasonable individual clinical and local decisions lead in aggregate to higher utilization rates, greater costs, and inadvertently worse outcomes.
Let's go to the next slide. You should see the question, "Why harm?" Right? At the bottom?
Male Voice: Yes.
DR. FISHER : So what I think are the three major causes—and these are outlined in the article that I believe you have about avoiding the unintended consequences of growth in medical care. There are three major mechanisms. In higher-spending regions we are more likely because there are more of us to treat patients with milder disease. We look harder for it. We diagnose more patients who would not have been diagnosed in the absence of those diagnostic tests. This is the problem of overdiagnosis. And when you intervene in patients who are not very sick, the benefits are quite, quite small but the potential for harm is almost as great as it is in the higher-risk populations, and this has been well shown in a number of studies.
Second is because there are more physicians involved with a given patient's care, the care is much more complex. It is less likely to be clear which physician will be responsible for following up with a patient about adherence with medications or for writing the discharge orders for their beta blocker if they have heart disease or for their cholesterol-lowering drug. Greater complexity of care. And what we see when look across either hospitals or regions, we see two- to three-fold differences in the proportion of patients who are seeing ten or more different physicians. You're three times as likely to have ten or more physicians involved in your care in a high-spending system, academic medical center, than in a low-intensity academic medical center.
And, finally, patients spend more time in the hospital. Identical patients spend more time in the hospital in high-spending regions. And hospitals are dangerous places if you don't need to be there.
Now, this is a—I gave this theory, this theoretical argument, to the American College of Radiology about eight months ago, and I got quite a vehement e-mail response from one of the members of the college.
Let's go to the next slide. He sent me a brochure that is being sent around to urologists by Siemens. He questioned whether this was just about the gray areas. And I want to share this with you because I think it's a real challenge, really, to the professional integrity of our—the professional integrity of medicine.
So if you go to the next slide, Slide 42, I believe, on your slide set, it should say, "Just the Gray Areas" capitalized.
CHAIRMAN PELLEGRINO : You got it.
DR. FISHER : So this is the cover of the brochure that was sent, I think, to probably all urologists in the country. It offered a new quick-start package. The notion was if you buy one of these CT scanners you can do quite well in your practice.
Go to the next slide. It's laid out exactly how easy it is to profit from this, so that in Scenario C, where the practice orders ten procedures per day, which is not a very high rate for one of these CT scanners, the return on investment over 5 years is $2.1 million in net revenue to the practice—profit to the practice.
Now, in case you're not a person who likes numbers and needs implicit messages to make this absolutely clear, let's go to the next slide, Slide 45. It should be a picture of a very happy urologist who is now able to sit back and relax because Siemens is taking care of the problem and helping him step by step.
Let's go to the next slide. It should be Geoff Smith, the radiologist. He mailed back to me, which I now share with my medical students as a challenge to all of us. He said, "These marketing ploys are wildly successful across the entire country. Patients are viewed as the ball in a pinball machine, popped back and forth, ringing up profits, until finally they escape past the paddles and can no longer render income. I believe that the fingers controlling those paddles often use those gray areas of judgment as an excuse to shoot the patient back to the triple-score bumpers.
"Speaking just as some guy out in the boondocks, I can tell you that life's more like the Star Wars trilogy than one would guess. There's a dark side. Difficult to resist and only a very few are able to throw themselves over the precipice to escape its clutches once they are embroiled within."
I gave this talk as grand rounds at the University of Colorado-Denver just a few weeks ago, and three physicians in the relatively small grand rounds audience came up to me and said they used to be in private practice but came to academic medicine in a salaried practice because of exactly this problem, the sense that they had to corrupt their professional values in order to practice medicine in the current environment.
So let me try to summarize the second portion of my talk about what's going on with a cartoon. Let's go to the next slide. If we're lucky, this should be a cartoon, two physicians standing side by side. Do you have it there?
CHAIRMAN PELLEGRINO : Yes.
DR. FISHER : "There. There it is again, the invisible hand of the marketplace giving us the finger."
So let's go to the next summary slide of the major points. "Overuse is largely a consequence of reasonable differences in clinical judgment that emerge in response to local organizational attributes (capacity, clinical culture) and financial incentives that promote unnecessary growth and more care." So that's the problem.
Slide 40, the next slide, should be setting us up, returning to the outline. The challenge we have is something about aligning incentives. The current incentives are not aligned with the best interests of patients or, I think, the best interests of physicians or the scientific integrity of our academic enterprise.
So do you have this outline slide back in front of you now?
CHAIRMAN PELLEGRINO : Yes.
DR. FISHER : All right. I am just wrapping up, so I'm going to summarize with what I think—I'd like to leave you with some ideas to consider. Let's go to the next slide. We have our group here. I've been trying to think of what are the underlying causes, the diagnosis, of the problems we see in medicine right now and what are the principles we should use as we think about trying to address the problems we see.
The first problem is I think there is a constrained view of science in academic medicine and public policy. We tend to focus on biotechnology. We tend to assume that the only problem is about correctly allocating the biomedical interventions, the specific treatments that we develop, and making sure they get to the right patient and that we have better evidence about it. But I think the data that we've developed over the last 20 years highlights the problems of care delivery in clinical practice, and we did not have good science understanding either the causes of the variations across U. S. regions or the consequences of those variations and what we should do about them.
We have been trying to do this, thankfully, with wonderful support from the National Institute of Aging and others, but this has not been a major focus of investment in the scientific enterprise within the United States . So our prescription would be moving toward a more balanced investment in the science of disease biology and the sciences of clinical practice and evaluation.
The second problem is our failure to recognize the key role of the local system, the capacity and clinical culture, as a driver of rising costs and poor quality and that that local culture happens at the level of the hospital and the physician group, the network within which physicians are practicing and the capacity of that local delivery system.
It happens across the academic medical centers that I showed you in that fifth slide, and I'll come back to those academic medical centers in a second. But the prescription here should be to foster the development of local organization, local delivery system, that can be accountable for the care for the populations they serve and take responsibility for the care of patients over time and that are large enough to have their performance measured.
We are working very hard with the Medicare Payment Advisory Council, with the Congressional Budget Office, Mark McClellan and I and others, to try to figure out how can we foster the development of accountable care organization that can take responsibility for coordinating care and that can have incentives not to purchase that MRI scanner, to be rewarded for conservative practice.
So the third problem is this assumption that more is better and the tendency that the American public has to equate less care with rationing. Now, if we look across the U.S. health care system, we see that the lower-spending regions are not rationing care, they are not denying beneficial treatment for the populations they are serving. The extra visits provided in the high-spending regions are not providing benefits on average. Patients might think they are, and that's part of the problem.
So the remedies here would be compared to the effectiveness research, much better information about the risks and benefits of the biomedical interventions that are under development and that we are developing to know in whom they are beneficial and what are the benefits and risks.
We then move to a model of informed patient choice, and we've been advocating this at Dartmouth for many years. Jack Wenberg has been taking the lead on this, and others, including Al Mulley , are talking about the fundamental importance of shifting from a doctrine of informed consent that essentially says, "I, the physician, knows what is the right things for you, and I think you should have this treatment." The doctrine of informed patient choice would say we have an obligation as physicians to present balanced information to the patients we see about the risks, benefits, and uncertainties and help them understand how their own values align with those different risks and benefits, and move toward a doctrine of informed patient choice: shared decision making where the patient's values are guiding the decision.
There is very important work going on in Washington state in this regard. The legislature there has passed legislation which will try to move the State of Washington toward a standard with the physicians of informed patient choice, which would protect the physicians from liability as well as provide much better information for patients about the risk and benefits, for example, of bypass surgery.
But the third thing we need in terms of the challenge of inadequate information and equating less care with rationing, we need much better performance measures about how local delivery systems are doing. I believe that if we could show information that was convincing and persuasive that a low-spending medical group or academic medical center that costs $3,000 less per beneficiary was providing actually better care convincingly to the public, the public would choose to have $3,000 in their pocket rather than receive unnecessary treatments, the wasteful care that characterizes the higher spending systems.
But that's going to be hard without addressing the fourth issue, which is our payment system rewards more care, increased capacity, high margin treatments, and entrepreneurial behavior. So we need to change our payment system, and I think we need to think about shared savings and some other models as an interim approach.
Let's go to the next slide. I want to end by highlighting the contrast between three members of the honor roll of academic medical centers and the differences in spending across these centers. You should have a slide that says, "Three members of U.S. News and World Report's Honor Roll": UCLA, Mass General, and the Mayo Clinic.
Medicare does a very good job measuring the quality of inpatient care and all three of these hospitals achieve high scores on these measures: 82 percent at UCLA, 86 percent at Mass General, and 90 percent at the Mayo Clinic in Rochester . But the differences in spending across these systems are fairly dramatic, nearly a two-fold difference. Medicare spending at UCLA on patients in their last six months of life is twice what it is at the Mayo Clinic, and this is a consequence of a pattern and practice that has 19 days in the hospital as compared to 13 at the Mayo Clinic and more than twice as many physician visits in the last six months of life at these higher spending systems. That greater resource use has very important implications for the future of the U.S. health economy.
Let's go to the next slide, and these are some questions that I'll leave you with that I think are challenging for all of us. What are the implications of these differences in resource use for the future affordability of health care? If everyone could adopt the current resource use levels of the Mayo Clinic, health care would be much more affordable. Might we be able to cover the uninsured? I think the challenges of covering the uninsured relate to an understanding of the challenge of the flexibility with which physicians and hospitals deploy themselves across the population.
Obviously if we can go from three to six visits per year for patients with well-controlled hypertension, we could currently accommodate the uninsured if we could figure out a mechanism for achieving coverage, and we don't necessarily need to increase spending to do so.
What about the challenges posed to the rest of the U.S. economy by the increased spending that is being devoted to medical care? And there are very different pathways offered by a practice that would look more like the Mayo Clinic than like UCLA.
Do we need to expand the physician workforce? Great question. If we adopted the physician workforce benchmarks at the Mayo Clinic, we would not need to increase the physician supply at all to accommodate the aging of the U.S. population and the baby boomers through the year 2020.
And what about the research priorities of academic medicine? We have not had much attention by academic medicine to the differences in practice across the hospitals, and I think that deserves some more attention.
It's really been a treat to talk with you. I'm sorry it was less of a conversation, but I needed to get some information across, and I'm happy to talk for as long as you are interested in taking questions.
We can go to the next slide if you like, which I think should be a reminder of what the health care system looks like when you look from 30,000 feet . We have some serious challenges. Thank you very much.
CHAIRMAN PELLEGRINO : Dr. Fisher , thank you very much for a splendid presentation, particularly under some trying and difficult circumstances. We appreciate it very, very much. Let me say it now: we hope you daughter is doing well, and we want to thank you for participating with us.
We will now turn the discussion over to members of the council, but first we have Dr. Daniel Fisher , who is going to open the discussion. Dr. Fisher, I'm sure you know, is Chairman of the Department of Medicine at the University of Texas Southwestern, an experienced clinician who will give us his point of view in opening questions for the rest of the audience. Dan ?
DR. FOSTER : Well, Dr. Fisher was the speaker, and Dr. Foster is the answerer, so—
DR. FISHER : Hello, Dr. Foster .
DR. FOSTER : Hi. How are you. I don't have a whole lot to say. I had read your paper and enjoyed very much the presentation of the slides. I think I would make a general philosophical statement to start with.
First off, from the early days of the country, we have operated in an economic system which is generically called capitalistic. The county has been filled with entrepreneurs. It looks like almost every place that you turn these days that the country is characterized by economic greed. What I mean by that is, we have major companies like Enron or whatever in which the salaries, even if you're fired for the stock going down, are enormous. I mean, the driving force—it's not alone for me to say this—appears to be economic greed.
It used to be when de Tocqueville was here, when he talked about communitarians, what he said is we actually have a tripartite system. We have a governmental system which is to assure the rights of citizens, including the very peculiar phrase in the Declaration, the right to pursue happiness. And then the economic system was designed to assure that there was enough in terms of income that one could achieve enough physical and other things to achieve happiness.
And then we had a communitarian system, which was quite remarkable. He thought this was the most remarkable thing, I believe, about the country; that is, that there were—all citizens who were communitarians, that they put the community above themselves, so they were volunteers. They supported hospitals and churches and museums and all the other things that fell through the crack between the government and the economic system.
And what we have seen, although there are companies that contribute massively to communitarian things, still there is this dominant sense of achievement in economic terms. It does not surprise me at all that if you have more physicians or more hospital beds in a community that the costs are higher because the hospital, like every company, and the physicians, like every worker, wishes to be economically secure. And as a consequence, they may from time to time or frequently choose between a lesser cost procedure or a higher cost procedure.
It doesn't surprise me at all that there is a relationship between the numbers of resources and the money that is spent. Moreover, I don't think it's likely that you can have a group of people to tell a hospital not to buy a 7.2 Tesla magnet as opposed to a 1.5 when you can make a fortune by doing cardiac MRIs. It's very helpful in the individual patient clinically. They're not going to do that because they're in the process of staying solvent and making money. And many of the systems—I can only speak about Dallas . The biggest private hospital in Dallas owns about 27 hospitals all around, maybe higher than that. And the drive there—and you get fired if you don't do it—is to make money or make your stock go up.
I don't believe that the drug companies are going to be humanitarian in terms—they will build—to get a patent difference that will allow them to maintain a drug will make slight changes that don't account for anything because they're in the business, as Relman pointed out, of making more money. Decisions are not made for communitarian reasons; they are made for profit reasons, because that's what our system does.
So I'm not surprised about that. Conversely, the reason we have no general internists is because they're so poorly paid. I have a son who is a general internist. He came to talk to me this week. He said, "Dad, I don't know how long I can keep going." He's an old-fashioned internist. He puts his own patients in the biggest private hospital and takes care of them rather than turn them over to a hospitalist. And we hire nurses that make more than he does. Out of our graduating class last year of residents—we have a very big program.
We have 150 residents in internal medicine, so about 55 go out. Three went into general internal medicine. And, as Relman pointed out when he was here, you can't get an internist in big cities. The only place that we have them is our medical school because we still take Medicare. Nobody else in Dallas takes Medicare as new patients. We don't do it because—if I see a patient, I may get $40, and I've spent an hour with a patient. And I write a lot of books and stuff like that. You would think in a merit system I ought to be paid more, but it doesn't happen. It doesn't happen that way.
So on both sides we see that it's the monetary factor that is making these decisions. And I believe—as Dr. Pauly said this morning, I don't believe we can change that. I don't believe that a group of saints could change that model. You might do it by explicit rationing, as we heard this morning, but nobody is—if you're Mother Teresa you're going to take care of people in India for free. You remember when the head of Time Magazine came there and she was scrubbing a gangrenous leg and he said, "I wouldn't do that for a million dollars." That's a very perceptive statement. "I wouldn't do that for a million dollars." And she said, "Neither would I."
I think the model is good. The only question that I would raise, and I may not understand the methodology that surprises me as somebody who has hung around with science for a long time, is that it was a universal conclusion of the study that all—all—of the top five communities and all of the lowest communities and little towns and so forth did better in terms of outcomes.
I'm very suspicious when I see ten or twelve points. All. I mean, maybe somebody in the third percentile or somebody—there surely should be at least some of these honorific academics, even if you give the beds and so forth. It looked too pure to me, but maybe I misunderstood that.
The final thing I'd want to say is that, in the last comments that were made, is that if a patient by choice chooses something—unless they're paying for their whole care—that's $3,000 less, that's not going to be money in their pocket. That's not the way it goes. And I think even in the honor place—I happened to have evaluated the Department of Medicine not too long ago at Mayo Clinic. I know that system extremely well. We spent—I spent a full week up there evaluating. They wanted to see how they were doing in terms of medicine.
The discrepancy between payments for physicians at the Mayo Clinic is different from what it is in a community like New York or Philadelphia and so forth. The salary discrepancies are not so much there. And, secondly, they've had that tradition for years, that you're a salaried physician. And it's cheaper to live in Rochester than it is to live in Los Angeles .
So my summary statement is that—and I want to say one other thing, and this I can't prove in any way at all except that I—it's a cynical thing to say that physicians make their choices for money. Now, if you said they make their choices solely for money, that would be even worse. But my own impression from interacting with hundreds of physicians is that most of them are decent, honorable people, and they want to make good money. And they go into cardiology because they get—plus, it's also exciting. But the implication of this just taken on the surface is that it's pretty sad.
I think the argument that Dr. Fisher is making is that it's not on the physicians. I think he's saying it's what he said several times. It's the community that you work in, it's the facilities that have to be paid for. if you can make more with more magnets and so forth. But it could be read that the physicians are cynical, and I don't want that to come away. I don't think he—I'll give him a chance to answer that. I'm talking more about that than I want to.
But I believe, to go back to what I said initially is, I don't think in this country that we can change the drive to make money whether it's at industry or whether it's in professional life you can do it. And that's why I don't know exactly what the solution would be to these things that we've seen.
So I enjoyed your talk very, very much, but it's—I'm not as optimistic. I think I would probably join with Dr. Pauly that I don't think that making saintly demands on a society is going to change things very much.
CHAIRMAN PELLEGRINO : Thank you very much, Dan . I hope you'll forgive me for the lapsus linguae . It's a combination of synaptical disconnect and perhaps, Paul, if you could explain the reason—
DR. FOSTER: Now, what I thought you were doing, I thought you were giving me a big compliment because Dr. Fisher had been so great that you would say, "Well, maybe one of these days you'll be a Dr. Fisher" or something.
CHAIRMAN PELLEGRINO : You're a very charitable man, Dan . Thank you very much. Now I will talk to Dr. Fisher . Do you wish to respond to Dr. Foster 's comments?
DR. FISHER : I would love to say a couple of things. One of the challenges is the mic—your end got cut up, and I may have missed a few things, but I want to comment on three things.
First, the issue of financial incentives and people needing to make money and greed; second, the issue of no findings that go better in the high-spending regions, and I'll come to that in a second; and then the question of whether there is anything to do about all this.
First, I agree with him completely that the United States has a culture of both hard work and wanting to be rewarded for one's work. And I agree with him that whether it's hospital boards—and I've spent a lot of time with them—or with physicians, I believe everybody is trying to do a good job. I do believe that the pressures on in the current payment system are outlined well in his comments about general internists and the impending collapse of primary care. The current incentives don't support the kind of systems that we need to provide good care.
The second question is, have we ever—sort of the skepticism that my findings were too cute. First I would encourage you to read the papers. We've really looked pretty hard, and we have one paper now under review looking at patient responses to care, patient perceptions of care at higher-spending regions compared to lower-spending regions in great detail, and there is in that paper one of the eight dimensions that we look at where patients prefer care in the higher-spending regions. So this is some good news there. And, interestingly, they like their doctors better.
I don't think I'm breaking some embargo of review and that my co-authors will shoot me. So there are some things that people appear to like better. But let's go to the third question of whether there is anything we can do about it.
My father, who is a lawyer, Harvard Law School , when asked about what he thought would happen in the future, he said that wasn't his job. His job was to try to make the future better. I think the question is, can we devise systems that would have incentives for good performance, incentives for better care coordination, incentives for physicians to work together, and allow them to be rewarded not for increasing capacity, not for seeing patients more frequently, but rewarded for good conservative practice.
And I think we're making headway. I don't—it's going to be hard work. The question of how we help physicians form groups that could be rewarded for conservative practice. It's a difficult one that we are working on at the current time. But I think if we designed—I am optimistic, and I've learned to distinguish hope from optimism. But I am indeed optimistic that we can make some changes to the current payment system that would allow groups of physicians and their affiliated institutions, to the extent we can foster collaboration rather than competition between hospitals and physicians, that would allow them to be rewarded for not growing at the rates they are now growing and were illustrated on the slide of Peter Orszag, head of the CBO.
The models are difficult to implement, this notion of a shared savings program. But I think we should look hard at trying to devise market mechanisms that would reward physicians both for good practice and for better care and would reward them financially. And that's the model that we're trying to work for.
I agree with you that simply exhortation, "Don't buy this scanner," that's never going to work. But I do think that the practices within systems with less capacity seem to be just as good as those with more, and the challenge is to get the incentives aligned to design good care-delivery systems.
So I appreciate all your comments. I think our challenge is how are we going to work together to do something about it.
CHAIRMAN PELLEGRINO : Dan , any further comments?
DR. FOSTER : This is Dan again. Thanks very much. I appreciate that. And I meant to say one other thing. I certainly agree with your first paper about the fact that getting all these CT scans looking for coronary in the arteries and so forth, oftentimes in the lungs you're going to see a little granuloma and then you have to take it out because you think it might be a lung cancer. So I just want to say that I want you to continue doing what you're doing. And I just wanted to add that little thing to it.
DR. FISHER : I appreciate that. This is hard work. We have a lot to do together.
CHAIRMAN PELLEGRINO : Thank you both. Are there further questions, comments? Dr. Rowley .
DR. ROWLEY: I'd like to know in discussions of how one moves forward and develops a more altruistic system, do you think that a single-payer system with the federal government is likely to be more conservative in terms of the use of medical facilities than the fee-for-service system that we currently have, or at least have in some areas and for some segments of society?
DR. FISHER : That's a tough question. It's a great one. I think that most visions of a single-payer system involve maintaining the fee-for-service approach. That is, physicians for a national health program single-payer system simply proposes to leave our current fee-for-service system in place but have it financed by a different tax system and payment system.
My sense is that will not solve the problems with the delivery systems that we've outlined. I think the question of whether a single-payer system or a multipayer system will be better for the care of patients is hard to answer. I think realistically politically we are going to end up with a multipayer system, at least over the next five years, and Leonard Schaeffer and others are predicting if we don't do something constructive to stem the growth of health care costs in the next five, maybe eight years before the next eight-year presidential cycle that there will then be draconian cuts just to payment levels, and the system will be severely punished. And we may end up with a single-payer system.
But I think multipayer systems might work just as well if we think about fostering this local collaborative model that's rewarded for better care. I've discussed this model now in Wisconsin and Vermont and New Jersey . Most recently it seems to be emerging in Oregon the notion that there has been a real barrier to physicians and hospitals working together to improve care and that we ought to—there's interest in models that would allow us to do better both in terms of care coordination across the spectrum of care that patients experience. And that's been the focus of my work. It's on the delivery system side, less on the payment system side. And I think we need to do both. We need to think about how to reform the insurance system, but my guess is we'll have multiple payers, multiple insurers, the government and others, and so we also need to work on the delivery system.
In Vermont the notion is a private-public collaborative to think about new payment models for these local delivery systems.
CHAIRMAN PELLEGRINO : Dr. Meilaender .
PROF. MEILAENDER : Thanks very much, Dr. Fisher . It was really fascinating. I just have one question about the general finding that in many respects the higher-spending regions have worse outcomes. I should maybe know the answer to this question. I don't know. But in your data, are these regions—there are undoubtedly variables among them with respect to population density, racial makeup, class makeup, and so forth. Is that all taken into account? Does that enter into your calculations in any way or not?
DR. FISHER : That's a great question. Our reviewers were quick to ask the question. If you look at that map one immediately thinks, hold it, Dr. Fisher is talking about the southeast, about New York and New Jersey compared to Oregon and Wisconsin , where everybody is healthy. The way we addressed that question was several fold. First, because we looked at patients with heart attacks or—just take heart attacks, for example. In those communities, the incidence of heart attack is actually lower in some of the low-spending regions, the healthier regions. But once you've had a heart attack, it sort of takes account of differences in the prevalence of risk factors across communities. That is, those who are heavier or have more risk factors for smoking are more likely to have a heart attack. But we looked quite closely, and the characteristics of the heart attack patients in low-spending and high-spending regions were very similar.
The second thing that we did was we said, "All right, yes, they are more urban." We repeated all of the studies, all of the mortality analyses, stratifying for different kinds of characteristics. So there are high-spending and low-spending rural regions. There are high-spending and low-spending urban regions. So we looked at this relationship in each of those strata, and we found exactly the same result.
I think we had 72 different strata we looked at: regions with low and high HMO prevalence; we looked only at blacks; we looked only at whites; men versus women; underlying risks based on predictive models of how likely you were to die following your heart attack or hip fracture. So we think we did a pretty good job taking account of those differences.
The review process at the Annals of Internal Medicine in their first articles list two years, eight different reviewers. And so I think the science is pretty good. I'd be happy to talk further, but we'd bore everybody to tears.
CHAIRMAN PELLEGRINO : Further questions? Anyone? Dr. Dresser .
PROF. DRESSER : Thank you. I'll take a shot at something. I wonder if you could put your findings into an ethical framework. It sounds to me as if you're saying patients in some regions are getting medically unnecessary care, and we know that because of the outcomes in the other regions. And so as a result, there's a lot of waste and some degree of harm to patients. And if we could take that waste and use it for other things, like covering uninsured, that would be an ethical improvement. Could you elaborate or correct me on that?
DR. FISHER : I think that's right. I think because we have focused our attention on whether a patient should get a very expensive drug or not and the likely cost implications or that or rationing. Because of the way we think about these things as clinicians, we focus on should you treat a patient with high cholesterol versus someone with low. We focus on the biology and intervention. We have missed this category of care called supply-sensitive services.
It's not about what care is provided, what specific services, it's about how the care is provided, which kind of provider delivers the service, what's the most efficient way. Should it be done in an outpatient setting or a hospital setting? We've ignored those. They certainly don't lend themselves easily to ethical analysis because those decisions are so difficult and require such judgment.
Do I admit this patient with heart failure to the hospital now or do I try to manage them myself overnight? There's no evidence to guide us on that. It's easier if I can get them a bed because they'll be hospitalized, and we think it's better care.
So I think there's an important dimension of practice within the United States that is ignored when we think about rationing care. I don't believe it is rationing not to provide services. Let me try to say that another way. I think rationing implies the denial of beneficial services, and so I think there is an ethical challenge posed by how to deal with this. I think when we talk about 47 million uninsured or we talk about starting to ration high cost but clearly beneficial treatment—a new drug is going to cost $100,000 a year but extend patients' lives substantially, and we say, "Well, maybe society can't afford it."
We ought to be thinking about the choice of providing that drug and all of the waste in current practice. I think there's a tremendous opportunity to improve the efficiency of care. If you take Salt Lake City as a benchmark, if you take—we're trying to do this at Dartmouth to reduce our utilization rates on these discretionary services. We recently decided, thanks to our data, not to build a new wing on the hospital. That will save the community a lot of money that we might be able to devote to other things. So I think there is an ethical framework here. Bringing it to bear in individual clinical encounter is going to be harder. And that's why I shift my focus for reform from an emphasis on individual clinical choices to context, environment, and incentives, focusing on the organizational context within which providers work and then trying to get the incentives aligned so that at least they are rewarded for providing better care even if it's less care.
One of the great challenges faced by organizations striving to do quality improvement work, say, now around heart failure or around other conditions, if you succeed in reducing—in improving care and lower hospitalization rates, you actually risk losing money. So we need to focus on the incentives.
Is that a partial answer to your question? I'm happy to follow up on the conversation.
PROF. DRESSER : Yeah, that's an answer to work with. Thank you.
CHAIRMAN PELLEGRINO : Other questions, responses, comments?
DR. FISHER : Can I make one more comment on the ethics?
CHAIRMAN PELLEGRINO: Yes, please.
DR. FISHER : I just finished an article. It will come out as part of the National Academy of Sciences' magazine, and I show that—basically I used the last slide. And I think academic medicine faces a serious, really ethical challenge. That is, how can we call for additional physicians, which we know will increase the cost of health care, and how can we look at the differences in practice and care delivery spending between—and you don't to choose just the Mayo Clinic. You can look at Intermountain Health Care. You can look within Massachusetts . You can look within New York City . Large differences in practice.
How can academic medicine—and is it moral? I'll ask you all to ignore the differences in spending and their implications for both the ability to pay for education for our children, the ability to cover the uninsured, the ability to pay for new clearly beneficial things while ignoring these differences in spending.
I think there is an ethical challenge there to academic medicine. At least that's the stance I'm taking, that my colleague Jack Wennberg and I are both taking. So help us with that. Is it an ethical question that we should pose to academic medicine?
CHAIRMAN PELLEGRINO : Dr. Hurlbut .
DR. HURLBUT : Maybe somebody wants to respond to what he just expressed first.
CHAIRMAN PELLEGRINO : I didn't see any—
DR. FISHER : It was a rhetorical question. Not to worry.
CHAIRMAN PELLEGRINO : There was no body language suggesting a desire to respond.
DR. HURLBUT : Dr. Fisher , I thought your presentation was really interesting, as were the readings you gave us. I want to ask you about something that we're going to talk about tomorrow morning, which is newborn screening. And, of course, implicit in that are larger questions about genetic screening at various stages of human life, preimplantation, antenatal, and postnatal.
Here's an example of something that relates to a particular paragraph in your paper which I had known about abstractly but was stunned to read, and that's the comment that it's increasingly clear that the population with an occult disease is many times larger than the population destined to become sick from it. And you go on to explain how microscopic examination of specimens from individuals without known cancer reveal high rates of thyroid, breast, and prostate cancer. I think my colleagues will know this. I put that on the record who are reading or listening otherwise.
As I thought about that and I thought about the implications of newborn screening, it really raised some interesting questions for me. I just thought maybe you'd have some things that could help us on this and it might relate actually into some of the other ideas you've presented to us. Four categories come to my mind, and maybe you could comment on each of them.
One is the cost of procedures like screenings themselves could exceed—could be very, very large and in many ways could displace the efforts we make to cure diseases. Do you see what I'm saying there?
DR. FISHER : Absolutely.
DR. HURLBUT: Then second, there's going to be a big problem associated with communication with patient populations that are not scientifically trained and don't necessarily have strong intuitive mathematical minds, and a lot of this information is going to just be statistical probabilities, not direct correlations between finding and the likelihood of disease. It's going to be a little like your paragraph: occult disease, if you will.
And then also it's going to invite all sorts of interventions, some of which will carry their own risks and some of them will have strange sort of psychosocial dimensions. I'm sure you know what I'm talking about with regard to that. And then finally the possibility that in the efforts to make a more cost-efficient system we may end up with impositions. In this case of newborn screening, they're mandated screenings now, which are kind of impositions, but for the most part we are pretty happy with that. But it could lead to impositions for interventions eventually—or at least strong pressures for interventions.
Could you just comment on this? We're searching in this council arena for ways we can contribute not so much to overall policy formation but with comment on professional conduct and ethical valences to this overall policy issue. So I thought this was a good subject you might be able to contribute to us in the perspective from which you've already delivered your comments.
DR. FISHER : I would be happy to talk about it, but the fourth point I didn't understand. I've got notes on the first three.
DR. HURLBUT : Well, it was a little more draconian, but the idea that once a genetic disease or proclivity to disease is identified, there might be strong pressure to intervene even when there is no symptomatology. You know, the preemptive things all the way from diets to maybe pharmaceuticals and so forth. I mean, here I'm thinking of—my mind goes to very troubling prospects, like suppose there's proclivities toward the use of addictive substances or antisocial behaviors or proclivities toward obesity and you could preempt that with a drug or—I'm just—you've spoken, I think, wisely and intuitively about the culture of practice that prevails. How are we going to make sure a culture of compassion accompanies that culture of practice when we come to the fact that many diseases are—many conditions are preclinical and yet we could intervene against them early. Do you see what I'm saying?
DR. FISHER : Yes. Now I understand completely. First, let me say that I think this is one of the most important challenges we face over the next decade. Because of advances both in imaging technologies and their sensitivity and in terms of our increasing ability to identify markers, be they genetic or others, that might be seen as risk factors for disease. And therefore we are tempted to screen and begin to intervene. So I think it's a critically important question.
I think the wisest person I know about this is the colleague with whom I wrote the paper that you're citing, Dr. Gilbert Welch, who has a book out about this, and it's the major focus of his work in trying to understand how best to help the public understand these issues and what we need to think about as we decide to implement new screening tests or to look for occult disease. You're absolutely right. Costs can be tremendous, and even the most recent example of the genetic screening tests for prostate cancer, which can stratify men into four risk groups. There's already the folks who published the New England Journal article who, I understand, are already offering a commercial test—little troubling in that on the evidence that intervening for those at low or high risk remains inadequate. There is no evidence with which to guide it. There has not been a study done yet to know whether the screening will be beneficial.
So the costs can be high and the consequences are uncertain and largely unevaluated. I think the challenge of communicating with patients and with physicians is difficult. We physicians are trained to believe that early intervention is better, and yet the science would say not always. So whether it's screening for lung cancer in the Mayo Clinic trial or earlier intervention in cancer therapy where stage shift is such a problem and—well, I think communication with patients and with physicians about the difficulty of understanding this problem is really important. And if you want, I can certainly help the Council by providing some of the papers on this issue for additional reading. So I think think you are absolutely right. Those first two issues—and the third issue about interventions having risks is absolutely correct.
Yesterday Dr. Welch was telling me about a patient followed for osteoporosis not having an easy time managing—she actually didn't have osteoporosis but was being aggressively treated because she'd been screened and found to have a slightly low T score. She had osteopenia, slightly thin bones. No intervention—none of the biochemical interventions seemed to help her, so they sent her to an endocrinologist who did a careful thyroid exam, found a little nodule. The nodule turned out to be papillary carcinoma, but we don't know what—most papillary carcinomas will never cause trouble, but she ended up with a thyroidectomy.
So you're absolutely right that these interventions have risks, and the public is largely unaware of them. Often they think, especially in the cases of early cancer diagnosis, that their life has been saved when, in fact, most of those cancers would not have caused trouble on average for the population If you follow them. This was overdiagnosing.
So I think in terms of the psychosocial consequences, we have to be aware that it cuts both ways. It could make people feel worse: "Oh, I have a cancer diagnosis. I've got cancer." But if you think of the vitality of breast cancer survivors and the sense of renewed vigor and renewed life that they get after they've survived their cancer, I'm not actually sure whether they would be assessed as being harmed or benefitted by their diagnosis, a very challenging evaluative problem for those of us who want to understand what is the impact of these diagnoses on population health.
I think the challenge of imposing treatments on populations—I think in all of these cases the question comes down as much as possible on demonstrating the benefits of intervention on an entire population while accounting for the risks of the harms of the intervention before we advocate adoption of the screening test.
So one thought—I mean, this may be an area—and there has been some talk in policy circles where it really would help the country if we paid attention to, say, the U.S. Preventive Services Task Force. What is the best evidence on this particular screening test? What is the quality of the evidence? And Medicare and other payers might wisely say, "We are not going to pay for and we discourage your receiving tests that are not recommended by a body that is evaluating their potential impact on population health.
Are those some helpful ideas? I'm not sure.
DR. HURLBUT : Yes, that was very helpful. I know there have been several studies that have recommended prenatal screening of not just women over 35 but of all women, and the costs associated with that will be huge. But what's troubling about that beyond the costs and what's seductive about it beyond the benefits of knowing something early and doing something positive about it, is that it puts the society at—it implies a certain attitude toward what it means to have a medical condition.
In other words, the very idea that you would know about having a disease either in the womb—a child who will have a disease or a person who has an occult expression before there is any manifest symptomatology, it creates a culture with a certain difference of attitude than the one that most of us grew up in. And I find it troubling because I think that—as I look out at it, I can imagine there would be some very significant cost savings associated with some of these early detection systems. Now, it could just be very burdensome and we decide not to do it, both in costs and in social process, but it could also be very efficient to do it, at least in some disorders, some of which might be actually controversial. Certainly prenatal screening is controversial because of its association with abortion.
So I just—I mean this is sort of a vague reflection now on your papers, but it really struck me that sometimes more medicine is not just less efficient in terms of its costs and somewhat of a burden on the patient but can change the character of your general society too, just general attitude of what human life is and how it unfolds naturally. And it can also do something, I think, that you haven't exactly mentioned, but make us a society preoccupied with disease.
I mean, I think pregnancy has already been affected by this, the sense that you have to do a lot of things to make sure your baby isn't going to be damaged. And a lot of that is very good. Of course, ultrasound and so forth. There are some interventions that can be done. But you know what I'm saying? The general feeling that we'll become more preoccupied with disease and what might cause our death than we do actually living our lives. Well, it's an exaggerated statement.
DR. FISHER : I share your concern. Both Dr. Welch and I and our colleagues up here at Dartmouth are working very hard to try to think about how to reverse the trend toward medicalization and how to help the public understand the choices they face. It's an uphill battle in a culture where, as I believe we discussed earlier, there's lots of money to be made by making healthy people sick. The device manufacturers, those who produce the tests, and those—the industry interest in taking treatments that are very beneficial in those who are very sick and the very small benefits in those who are pretty healthy.
But the goal of industry and the pharmaceutical industry, because there is such a large population there, is to make us—is to shift the definition of disease to increase the population. And Dr. Welch—I almost included three slides from Dr. Welch's recent synthesis of the work on hypertension where it is clear that we are labeling more and more patients what used to be prehypertension and no one would have mentioned it to you except maybe tell you, "Dr. Fisher, you ought to do a little more walking" or "Why don't you watch our diet." No one would have mentioned the diagnosis, but now we are being encouraged to treat vast numbers of people.
Most of the population has mildly elevated blood pressure. Two things happen. One is you're treating people who have a very small chance of benefit. You're labeling them as hypertensive, which we know causes increased absenteeism in the studies that have been done, but we also are more likely, because there are so many of those patients, to ignore the patients with severe hypertension. Right now only half the patients with established hypertension on average are being effectively managed. So not only do we risk labeling the less ill people, but it diverts all of our attention from those who really need our care and would benefit substantially from better care.
So I think you're posing a great question. I'm not sure I can solve it for you for the prenatal screening, but I think it is a very important problem for society to come to grips with, because otherwise we will all be sick.
CHAIRMAN PELLEGRINO : Dr. Peter Lawler .
PROF. LAWLER : I really am listening to all this from the point of view of someone who doesn't know much about it. I'm not a physician. Except I really don't know what to think about it. I mean, Bill is right. We live in a time when people are preoccupied with risk factors of every thought, every kind, and prehypertension just sounds dumb. I mean, next it will be pre-prehypertension or something. And so, no doubt, you are getting people alerted unnecessarily and so forth. But the general tenor of the discussion is something like this: if medicine is too conservative, your cancer will be diagnosed too late and so you will die. If medicine is too aggressive, you'll be treated for cancers that you never would have known about and never would have bothered you, and you'll have a thyroid procedure that is completely unnecessary.
So what's an average guy to do here? We don't want to be too conservative, but the reigning wisdom is we don't want to be too aggressive; we don't want to be too conservative. How can anyone measure exactly what is correct on this? Because on the one hand we don't want to get all excited about prehypertension or pre-prehypertension; on the other hand, perhaps we're not sufficiently concerned about real, life-threatening hypertension that's a big-time risk factor.
So in a certain way, doctor, aren't you just pushing in the other direction against aggressiveness, telling us in fact that conservative medicine is in fact less dangerous. So you may be convincing me that I should never go to the doctor again for fear of what might happen to me. On the other hand, when people die of cancer you always get that "Tsk tsk. Went to the doctor too late. Weren't aggressive enough."
CHAIRMAN PELLEGRINO : Thank you, Peter . Dr. Gazzaniga .
DR. GAZZANIGA : Elliott , nice to hear your voice again.
DR. FISHER : Yeah, hi. How are you doing?
DR. GAZZANIGA : I'm fine. I was wondering if you might comment on the general question of group versus individual data. As we all know, there's tremendous individual variation in biology and biomedical phenomena, and so much of medicine now is really trying to aggregate groups, find the mean and see if there is a significant difference between them when within each group there's tremendous individual variation. So the drug action on you may be one thing; on me it may be another. And yet if we put this into an average it looks like there may be no effect of the drug.
In our own MRI work we've always moved forward by group data, but now we're looking at the individual data. It's so reliably specific and varied that you have to take note of it. And also as we're moving biomedically into more individual medicine, personal specific biologic state medicine, how are you going to deal with that, or how is biomedicine going to deal with the fact that the individual variation is really where modern medicine is going?
DR. FISHER : Well, thanks for giving me such an easy question. This is a really important problem. If you look at Rodney Hayward 's work, which I'm not sure many of the council members will be aware of, but a very nice set of studies from a guy at the University of Michigan who sort of highlighted the problems in randomized controlled trials and the use of the group data. What his work has shown and some have shown is that you can have two different problems. One is interventions that look effective on average can be highly beneficial to some in the group and either of no benefit or harmful to others. And if you have a negative randomized trial that shows the drug didn't work on average, you can then look at the data and find in fact there were subgroups who benefitted from the drug and others who did not. Therefore, the average effect was null or negative but there was a subgroup who benefitted substantially.
The Institute of Medicine held a workshop last summer of which this was one of the problems discussed. And I think we're compelled as scientists to try to learn about whether the individual predictions we make based on our understanding of biology and our understanding of the pharmacology and pharmacogenetics, if you will, of the patient and try to test our prediction that these patients will be likely to benefit. So the notion of much more comprehensive evaluation of drugs, much more effort to prespecify subgroups early that might benefit, much greater use of observational research with good baseline data collected on the characteristics of the patients that would let us understand better how the heterogeneity within the population affects the benefits of the treatments that we're providing patients. And it's going to require what the IOM called a learning health care system.
I think the emergence of good electronic health records, the emergence of some sort of population-based registries that allow us to follow patients over time and do much better postmarketing surveillance of both devices and new pharmaceutical agents offers at least the best shot most smart people—and I probably wouldn't include myself in those—the best shot at addressing the question you're asking. Is that a fair—it's an important problem.
DR. GAZZANIGA : Yes. Thank you.
CHAIRMAN PELLEGRINO : Further questions or comments? Dr. Fisher , any closing remarks?
DR. FISHER : I want to thank you all for listening to a disembodied voice. I know it can be very hard. I really appreciated the opportunity to be with you and your flexibility in acknowledging my need to be with my daughter this morning. She is doing well. So thank you very, very much, and please let me know, Dr. Pellegrino , if there's any additional materials that we can help the Council with.
CHAIRMAN PELLEGRINO : Thank you very much. We really appreciate your being with us. The circumstances under which you've done it are a bit difficult. We understand that and are glad to have the news about your daughter. And on behalf of all the members of the Council, I thank you.
[1 ] Was unable to identify who said this.