Intro. [Recording date: July 26, 2018.]
Russ Roberts: My guest is Dr. David Meltzer.... I found out about you from a New York Times Sunday Magazine article by Kim Tingley in May of 2018 that was extremely provocative, and we'll link to that article. It's about the importance of the doctor-patient relationship and continuity of care, the importance of patients being seen regularly by the same doctor even when they were in the hospital in something of a crisis situation, an acute situation. And the article mentioned your attempts to actually quantify and measure the benefits of letting patients continue to see their own doctor once they are in a hospital. Now, as background for understanding your research, I want to back up and talk as you do in your work about the phenomenon of how hospitals have changed and what a hospitalist is, and how that has changed in recent years in terms of hospital practice. So, give us some of that background, please.
David Meltzer: Sure. So, hospitalists are doctors who specialize in in-patient care, and with some idea that, got that name in the mid-1990s with an article published by Bob Wachter and Lee Goldman in New England Journal of Medicine. And, the hospitalist model is a change from what's been the historical model in the United States for general medical care where people generally had a primary care doctor who would see them in clinic for most of their needs over time; and then when they got sick enough that they needed to be hospitalized, that primary care doctor would also care for them in the hospital. Traditionally the way primary care doctors did this is they reserved their mornings or a part of their mornings to see patients in the hospital; and they'd see their own patients in the hospital during that time. And then when they were done seeing those patients they'd go over to their office and see patients in the afternoon in their office. And that changed with the growth of the hospitalist movement, where the hospital caring increasingly provided by a different doctor, a doctor who just saw patients in the hospital. When that model got started, the argument one heard primarily about why it was growing is that hospitalists would do a better job: that, by being in the hospital all the time, they could be more present to provide the care that patients really needed. And, they would gain expertise in it because they did more of that. And so--
Russ Roberts: They'd save travel time, obviously. There would be--there's less of a burden on the primary care physician of the patient.
David Meltzer: Well, that's true. But that's not what people said about it. What people said at the time was really--
Russ Roberts: Specialization.
David Meltzer: Yeah. Specialization. They do more of it, therefore they do it better; they're going to get better outcomes. That sort of thing. And that hypothesis was studied, and I would say, over time there's been some evidence assembled that hospitalists can decrease length of stay a little bit; maybe improve outcomes a little bit. But I would say that the big picture of that literature is that hospitalists are not game changers. Not producing fundamental changes in hospital outcomes or resource utilization. Although I do think there are some benefits. And this sort of absence of absolutely compelling benefits made me wonder why hospitalists really grew: was it maybe there were such benefits; maybe people believed there were such benefits and adopted it for that reason. That certainly may be the case. But, there was another theory that sort of came to me as I was thinking about this from an economic perspective--and you sort of started to allude to it with your comments. And the theory was that hospitalists not really because it was better for patients per se but just because the old primary care model declined not because it wasn't better for patients but because it just was no longer economical for doctors to reserve their mornings to see patients in the hospital. And the reason for that wasn't that there weren't just as many patients being hospitalized, but instead that the people who were coming to primary care clinics in general were healthier and healthier. And so, those doctors could spend all day in clinics seeing lots and lots of patients and be very busy, but almost never have anyone in the hospital. And so, the increasing ambulatory volume compared to sort of stable hospital volume really pushed primary care doctors to take up, in a sense, the offer of other doctors willing to provide hospital care to take that off of their hands. So, that's what we call the ambulatory economics model of hospitalist growth. And we wrote an economics working paper about it and actually tested the theory with data on which primary care doctors decided to give up their patients to hospitalists. And what we found is that sort of all the different elements of our theory were supported: That, primary care doctors were more likely to give up their patients to hospitalists as they had a lower likelihood of having each patient admitted to the hospital, as they worked fewer hours during the day; as it became easier to just use technology to just call the hospital as opposed to going on yourself. And also as transportation costs rose: traffic increased, hospitals got bigger, parking lots got bigger--all that sort of stuff. And so, from that research we came to the conclusion that hospitalists grew really not because it was better for patients, perhaps, as much that it was just more expeditious for primary care doctors to turn over the hospitalist work to hospitalists.
Russ Roberts: So, in a normal market, the benefits to the primary care physicians through competition might be passed on to the customer in the form of lower prices, so that doctors really found it less efficient to make those trips. Maybe they didn't get to see their patients; that, through competition, would share some of the savings with their patients. Of course, that's not the way our health care market works. So, it just kind of happened.
David Meltzer: Yeah. And I think--one thing I'll just point out is that if you look at things like concierge medicine which are very market driven approaches to medical care, it's not uncommon at all for the concierge physician to do exactly the old job--that they'll see you in clinic; they'll see you in the hospital. They'll see you wherever it is you want them to. At home, almost.
Russ Roberts: Explain, for people who don't know, explain--someone might think a concierge physician is somebody who you find at the hotel. But, talk about what that model is for people who haven't experienced it.
David Meltzer: It's also sometimes called direct primary care; and I think there are a couple of different flavors of it. But, I would say that the gist of it is that the patient is paying out of pocket for the doctor's care. They often are paying a monthly retainer. It may well be that, having paid that retainer, there's no fee for additional care; or there may be fees in addition based on the care provided. But I would say that the general idea is that these doctors are charging higher prices, typically directly to the patient; and in exchange for that providing greater degrees of access.
Russ Roberts: They are also coordinating care when there's a complication that a concierge doctor doesn't know how to deal with--a specialist, right? They are going to find you the specialist, help you get access to the specialist--
David Meltzer: Sure--
Russ Roberts: and as you say, keep an eye on you through the process.
David Meltzer: Yeah. Yeah. They hopefully are arguing whatever you need--that is, direct care themselves, or facilitating connections to other doctors or helping connect your care team. Basically, what you need is what they hope to give you.
Russ Roberts: You know, David, I like to joke, when people ask if I'm a doctor, I say I am--because I have a Ph.D. in economics--I say, 'I am, but not the kind that helps people.' Which is not my joke: it's a friend of mine's. But I will say that I have a couple of friends who, I have the concierge relationship with. I don't pay them: they are just buddies of mine in my synagogue where, when I am anxious about something, I say, 'Is this a problem?' and they say, 'No, you're fine.' And that is an immensely valuable service that I assume a concierge doctor is actually doing.
David Meltzer: Oh, yeah.
Russ Roberts: For my friends--I want to say, 'You know, I could do that, too.' That's the level of medicine that I am capable of delivering: 'Oh, no; you're fine.' But I do think it's interesting how eager people are to be reassured that it's nothing serious.
David Meltzer: Sure. Sure. But I'm not sure a good concierge physician would just reassure you all the time.
Russ Roberts: I get what I pay for. These are unpaid concierges. They are my buddies.
David Meltzer: Yeah. Yeah. Well, anyway--I think the point is, sustained attention is a good thing; and sustained attention from somebody who actually knows what they are doing is even better.
Russ Roberts: Yeah. Good point. I accept that.
Russ Roberts: So, carry on. You started to wonder whether this system was--what its implications were and whether this loss of continuity was meaningful in terms of outcomes.
David Meltzer: Yeah. Yeah. I mean, intuitively, what is it that you give up? You give up the continuing relationship between the doctor and the patient. That has a bunch of elements to it. There's the knowledge that the doctor and the patient have of each other, including of the medical history and their interactions with each other. There's hopefully their ability to communicate better. There's their trust--which comes partially from just knowing each other but partially from the continuing relationship and the understanding that you are going to see each other again and again and again. And there's interpersonal relationship, which doesn't always get better when you know someone for longer. But it's getting worse--it's presumably going to cause you to change; and eventually gravitate toward someone with whom you really have that. So, there are all those elements of the doctor-patient relationship. I'm sure there are others. Those certainly have been documented well in the literature. There's also a wonderful observational quantitative literature showing things like lower health care costs when you've had the same doctor for a long period of time; the avoidance of unnecessary care at the end of life when you are cared for by a doctor who really knows you. Those are observational studies. There are even a few experimental studies where people were randomized or effectively randomized to have continuity or discontinuity in the relationship with their doctor. And in those studies, there's very clear evidence that that continuity has really improved outcomes, decreased hospitalization, and so on. So, there's lots of reasons to believe that this disruption of in-patient and out-patient, between in-patient and out-patient care could be costly. And that, I think, was well understood early on by the leaders, the early leaders of the hospitalist movement. Accordingly, one of the consequences of the hospitalist movement has been a greater appreciation of the need to improve handoffs and of strategies for doing that. But, you know, taking a problem and trying to reduce its harm is not the same thing as eliminating the problem, both because you often don't completely eliminate those things and also because the interventions themselves are often costly. And, if you look at the literature that looks at care coordination interventions, you find some work; some don't work so well. It's extremely difficult, perhaps impossible, to point to an intervention that consistently mitigates or eliminates the problems of discontinuity and actually succeeds in saving money. So, there is a real challenge that comes from discontinuity and thinking about ways to mitigate it is an important problem.
Russ Roberts: An EconTalk listener who is a physician wrote me earlier this week or last week--I can't remember--and mentioned his hatred of electronic medical records. And in his case, what he disliked was how it changed his interaction with the patient. Instead of looking at the patient's eyes, his face, body language, he found himself looking at a screen. It tended to--I think the way I would put it--is narrow the level of responses, information he was keeping about the patient, because things tended to fill in a form rather than be more freestyle. But, ironically, what struck me when he wrote me is that one of the ideas of electronic medical records is to improve continuity. The idea would be that a new doctor--true, they haven't chatted about their kids maybe with their patient and the doctor, but you'd at least have "all the information." You'd know things about--to be trivial--their allergies, their medications, possible treatments they were undergoing; they might interact negatively with things you might propose. And so, in theory, the whole idea of the electronic medical record is to help make that handoff more effective. What are your thoughts on that?
David Meltzer: I think all those things are true, to some degree, and the real question is how you weigh them against each other in various contexts. Whether you are a doctor or a patient, if you've been in a clinic recently you've experienced the changes that take place with the presence of an electronic health record. There's more information more readily available, but it requires the attention to some and sometimes to a lot of degree of the clinician. That means more of your time is spent looking at the screen and less of your time is spent looking at the person. I think doctors are aware of this challenge; and I suspect the vast majority try to find a good balance between the two. Some clinic rooms are better set up than others so that you can look at the screen and look at the person. But, truth be told, electronic records, health records, typically still take more time to complete the documentation of the visit than did the old of scribbling something on a piece of paper. There are surely some advantages of that: It's easier to find allergies; the previous medical history is probably recorded perhaps in a more organized way. But, the electronic health record can be, you know, thousands of screens long, and there's no time to look through all of them. And, so information gets lost. And it takes a lot of time for the doctor and out of the encounter. Which means that other things presumably don't happen. Including, you know, looking the patient in the eye and noticing that their facial expression is just a little different than it might be otherwise. Or, the sort of awkward pause at the end of a visit when the doctor thinks the visit is done and somehow the patient isn't quite getting up--
Russ Roberts: And they're afraid there is something they are afraid of, or embarrassed--
David Meltzer: Yeah, there is something else to be said that hadn't been said. So, you know, over time presumably, electronic health records will get more efficient. We'll learn what's more valuable and what's less valuable. But, these are not costless interventions. I think there was a really interesting point made in the New York Times article that I thought was really, really thoughtful. And, what she highlighted in it was that electronic health records are in a way a tool that are designed to promote communication between doctors. And when you require the adoption of electronic health records, we are requiring something for which the purpose is fundamentally to lower the cost of specialization and division. As opposed to, for example, lowering the cost of continuity, or promoting continuity and relationships. And, I mean, it's a really interesting thought--it reminds me of the debate about the subsidies for public transportation. Why is it that the riders of public transportation shouldn't bear all the costs of that? Why should there be a public subsidy there? And, you know, the interesting counterpoint that of course: The costs of roads aren't really paid by drivers. Right? And so, it's a great reminder that when we make policy decisions, we may perhaps not explicitly but implicitly subsidize certain ways of organizing care. And I think there is a significant degree to which our decision to subsidize the electronic health record, to require that, is predicated on a belief that, sort of, specialization is valuable, and that medicine can be broken down into algorithms; that there are formulas for these things. And, I fully believe that there are measurable things that electronic health records can help us with. But, I also believe there are things that we are a very, very, far away from having medical records really help us with. Like, understanding the mood and the affect of the other human being that we're sitting there in the room with.
Russ Roberts: Yeah. It's an incredible example of the complexity and the subtle and non-measurable things that can make a difference, which really I see your work as being part of. This piece of the conversation is related to an earlier episode we did with Jerry Muller on the Tyranny of Metrics and how our desire to measure things--a good desire--leads sometimes to a form of what Hayek called scientism. It's really fake science. It looks scientific. And I also assume there's a temptation for there to be a creep in the scope of what's gathered in the electronic health record--that it doesn't stay static and necessarily doesn't always expand in the direction that's always desirable. But it probably doesn't stay still. And more time gets taken up that this or that will lead to a better outcome--but might not.
David Meltzer: And--but let me sort of anticipate a place where we'll head at some point. I think it's a good moment to talk about it. Which is that it's a fact that the vast majority of money in health care is spent on a very small fraction of the population--
Russ Roberts: 80:20, roughly?
David Meltzer: Um, different measures, different things. But, yeah, something like that is about right. And I go beyond health care expenditures to say that the vast majority of us are pretty well the vast majority of the time. And yet, there are moments for all of us, and often extended moments for some of us, where people are really sick, and have really complex care. And often what makes that care complex is the interaction of myriad forces. It's not just that you have heart disease, but you have lung disease; and you are depressed; and you've just gotten divorced; and you lost your insurance. It's a million interacting things that are incredibly difficult to put in formulas. And so, what do electronic health records do? They are typically pretty good at taking formulaic approaches to fairly straightforward things. And, that's not a completely fair characterization, but I would argue that sort of, by and large, the more complex things get, the harder it is for electronic records to really make them formulaic. So, the consequences of this, ultimately, is that when you prioritize the things that electronic health records can do, you often prioritize things that are not necessarily geared toward where the greatest human suffering is, where the greatest costs are. That's not to mean that isn't [?] to get a Pap smear, or a mammogram. But, those decisions are very unlikely to be the fundamental drivers of health outcomes and costs.
Russ Roberts: You look for the lost keys under the lamppost. We won't go into this, but you'll appreciate this: It reminds me of a macro model of the economy: 'With just a few simple equations we can capture most of it.' Of course, you can; especially when things are going well. The irony of it is, it's really good at predicting interest rates when the economy tomorrow is going to be like it was yesterday. When it's not, it doesn't work so well; and we don't really understand the complexity fully, or the causal relationships. The body reminds me a lot of the economy--as listeners know.
Russ Roberts: So, you then decided--I think there had been one other reference that you mentioned from the Veterans Administration--but you had this vision--which is crazy--of trying to create a randomized control trial: a full-fledged experiment that would allow a comparison between patients who have continuous care who see the same doctor once here in the hospital, versus ones who don't. And these are high-risk patients. So, talk about the population and how you designed the study.
David Meltzer: Yeah. Well, even before that, let me sort of describe a little bit the rationale behind the model. We've already talked about this idea that continuity is a good thing. But, as I mentioned, a problem that we had recognized was that it just wasn't practical for primary doctors to do this old job, because they didn't have enough patients in the hospital on a daily basis. So, as we came to understand that, what we realized was that there was a solution. And, the solution was that if we could increase the rate at which patients were hospitalized for a doctor, it might still make sense for them to do that old job, providing continuity in and out of the hospital. And, of course, you can't just do that by increasing the rate of hospitalization to entertain doctors. That would not be a recipe for good or efficient care.
Russ Roberts: Or poisoning patients to increase their hospital visits.
David Meltzer: Right. Bad idea. So, but what we realized is that you could prospectively identify those patients who were likely to be hospitalized. And then have a set of doctors who focus their practice only on patients at high risk of hospitalization, so they can have a panel[?] size--a group of patients for whom they were responsible--that was small enough that they could provide them with all the ambulatory care they needed while only having clinic in the afternoons; and yet have those patients be in the hospital enough that it made sense for the doctor to reserve their mornings to see those patients in the hospital. So, basically, you could have a set of doctors who practiced medicine in the old way and have continuity in and out of the hospital, and for whom it's economically logical to continue to do that, because they only see patients at high risk of hospitalization.
Russ Roberts: That's very clever.
David Meltzer: Yeah. It was a really neat idea that came straight out of good old microeconomic theory, essentially examining the budget constraints or time budget constraints: the more doctors practice in the old model of in-patient and out-patient medicine combined and doctors dividing up tasks.
Russ Roberts: Did you have to have to recruit the doctors to be willing to do this in this unusual way?
David Meltzer: Yes. So, what happened is, I wrote the economic theory paper, came up with the idea. And then the Affordable Care Act is passed.
Russ Roberts: Which you were not consulted on. Go figure.
David Meltzer: Not in particular. But, as part of it, they created the Center for Medicare and Medicaid Innovation [CMMI], which was supposed to figure out what works in health care: How do you actually make it better? And, I had an experience both as an economist and researcher but also as a physician, and a physician who had taken on political leadership responsibilities of running our hospitalists' group; and so I thought it might be possible to hire a bunch of doctors, connect them with our hospitalists' group, and actually build this model at the University of Chicago. So, when the Center for Medicare & Medicaid Innovation was created and they had their first round of innovation awards, we applied, and we were very fortunate to be funded and given money to actually start this program. At the U. of C.--the University of Chicago--I hired 5 doctors who would do this, to gradually building up the group from 2 to 3 to 4 to 5. And CMMI allowed us to establish this intervention and also to do a randomized trial as part of that. So we had, in the end, about 2000 patients, 1000 in the intervention group and 1000 in the control group. And that started in November of 2012 and took us[?] until I think June of 2016 to randomize the 2000 patients; and we're now in the process of analyzing the data from following those patients, now over at least a year but in some cases now several years.
Russ Roberts: What do you mean by it took 4 years to "randomize the patients"?
David Meltzer: Well, we had to find people who were willing to be the study. It was a randomized trial; we sent[?] people and asked their approval before they enter it. In order to enter the study, they had to meet certain criteria, like they had to be insured by Medicare; they had to have been hospitalized once in the past year--which is how we identified a group of people who were at high enough risk of hospitalization or have some other markers that suggested they were at high risk, or higher risk, of hospitalization. And, they had to be willing in principle to give up their primary care doctor, if they were randomized in the intervention, because otherwise it wouldn't have made sense to include them. For the patients who were randomized the intervention, they were given the option to take that [?] comprehensive care physician as[?] their primary care doctor. For the patients who were randomized not to have that option of having their comprehensive care physician, we offered them help to find a new doctor. Because we didn't want to just compare our outcomes, our patient experience to the outcomes and patient experience of people who were unhappy with their doctor. That wouldn't have been a very interesting study.
Russ Roberts: So, I'm confused. We've got 2000 patients. We're going to put 1000, we hope, in the innovation group and 1000 in the control group. The innovation group, I assume are going to have the continuity.
David Meltzer: Right.
Russ Roberts: I don't understand what it means that they had to give up their primary care doctor. I thought their primary care doctor was going to be their continuity doctor. Explain how it worked.
David Meltzer: No. So, example: You come to the emergency room, or you come into the hospital, and you are admitted. And we approach you and say, 'We have this new program where you can have a primary care doctor who will care for you both in clinic and in the hospital. We are studying it through a study, so we can't guarantee that if you are interested in joining this study you will get a doctor who will do that. But you'll have a 50% chance of getting a doctor like that'--
Russ Roberts: I get it. And that's going to be different than their regular primary care doctor--that they've had so far.
David Meltzer: Right. If they've even had one. Some people won't have had one. But if they have had one, and they are willing to enter the study, you might imagine that they are not very happy with that person.
Russ Roberts: Good point.
David Meltzer: And so, we don't want to compare our outcomes to--
Russ Roberts: unhappy ones[?]--
David Meltzer: the outcomes of patients who are unhappy with their doctors. So, we tell the patient that if you are not randomized to the intervention group, so that instead you are randomized to the control group, you can keep your primary care doctor if you want; or, we'll help you find another one you'll hopefully be happy with.
Russ Roberts: Brilliant. Okay. Very clever.
So, that's what you did over the 2012-2016 period. And you got your sample. And now two years have passed, approximately, or a year and a bit. And, you've got some preliminary results that you're allowed to share. So tell us what you've found so far.
David Meltzer: Yeah. So, we're still doing the analyses, but we've presented some of the results publicly, and that's what I can talk about. We organized our evaluation around what's sometimes called the triple aim: sort of better patient experience, care; better health outcomes; and lower utilization and costs. And, for patient experience, the primary measure that we've looked at is how the patients rate the quality of care that they get from their doctor. And the bottom line is that people in the control group--well, people in both groups when they started were pretty unhappy with their doctors. Like a 20th percentile nationally. The people in the control group actually went out to about the 80th percentile. And we think that's because we helped them a doctor they liked--or at least didn't dislike. And then the intervention group went up to the 95th percentile in patient satisfaction. So, we thought that was a really big success, and we think that's important [?] patients.
Russ Roberts: Of course, one of the issues there, which I'm sure you worried about, is that when I'm rating my doctor who I've known for a while and it's the same doctor all the time, I might feel bad about saying something negative about him. So, you are asking them on a scale of--1 to 5? What's the--
David Meltzer: 1 to 10. And, the vast majority of people give their doctors 9 or 10. You know, even when they are not all that happy. So, there's a lot of top-coating, but nevertheless we found that our CCP[?Comprehensive Care Physician?] doctors did better.
Russ Roberts: CCP stands for?
David Meltzer: That's the Comprehensive[?] Care Physician--the continuity in and out of--
Russ Roberts: The continuity intervention.
David Meltzer: Exactly. And then, in terms of health status measures: we didn't find differences in general, self-rated health status. But we did find statistically significant improvements in self-rated mental health status. And lots of reasons to believe that that's related to these stronger relationships and that sort of thing. And then finally we looked at hospitalization, which is the biggest driver of costs in this population. And we found what looks like a 20% reduction in hospitalization rates after a year. Which is huge. And probably translates to savings of several thousands of dollars per patient per year, which is just an immensely large amount of savings.
Russ Roberts: Yeah. 20% of any medical cost is going to be a lot of money.
David Meltzer: Yeah. It's a lot of money. And, you know, I think what makes it really notable is that whereas a lot of the care coordination interventions that have been used in many instances require hiring a lot of people in order to bridge that care, our model really doesn't require that. In fact, all we're really doing is reorganizing care so that we don't need to have as many of these handoffs. And so, you could call this a lean approach to care coordination. It has gotten rid of potentially avoidable unnecessary costs of communication. And so, the fact that it looks like it's reducing hospitalizations a lot, and therefore presumably saving a fair bit of money: wow! It costs very, very little to run this program. And it's really just reorganizing care and getting patients into it. It really suggests the promise of this sort of approach. Not to the exclusion of other forms of care coordination; but I would argue as a strategy to lessen the need for them, and perhaps allow those people doing care coordination to focus on the areas where they were really required.
Russ Roberts: Quick question on the mental health: How is that measured? Self-reported, I assume?
David Meltzer: Yeah. We just have--in the results I'm describing to you, just sort of self-reported health status: How would you rate your mental health? Excellent, Very good, Good, Fair, Poor. A fairly crude measure but one that's been shown to correlate with other measures of mental health. And we have a lot of other measures of mental health in our data, but we haven't yet reported on them publicly. So, that will be something to look forward to.
Russ Roberts: Let me just ask a more general question about continuity of care and how it actually works in practice. So, my father last year was hospitalized for some surgery, cancer surgery. And I'm happy to report he's doing fine. He's on his fourth cancer. So far he's beaten all of them, which is glorious.
David Meltzer: Great. Good for him.
Russ Roberts: Wonderful. And I'm grateful for that. But I was with him in his room for, you know, hours at a time in the aftermath of the surgery, and one thing that struck me was how many doctors, and nurses, trooped through that room. Who were unbelievably specialized. And, to me, pretty ignorant of what was going on in the room. Other than: They would pick up the chart; there's a thing on the wall--the equivalent of a whiteboard with a bunch of scores of various kinds that could have been electronically recorded somewhere; I assume they were, but it was interesting that they were on a whiteboard with markers. And, some of them seemed great. I have no idea how competent any of them were, but they varied widely in their ability to reassure my dad that he was doing okay or things he was worried about were important. More importantly, reassure me, who was acting as his representative. How would that have--and I counted them at some point. I've forgotten the number. I was just shocked at how many there were. Same with the nurses, the people who gave him, you know, minute-to-minute comfort for things he was struggling with in the aftermath of the surgery. You know, there were nurses who did x, and y, and z, and tested him, and poked him, and brought him other things. And, it was just--it seemed like a recipe--and I've read this, of course--it seemed like a recipe for things to go wrong, just because of, just the sheer number of people that he was interacting with. And, how would that experience change in the story that you would imagine, in a different world?
David Meltzer: Yeah. So, let's take your dad's case very precisely. And he's got a malignancy and in the hospital. So, in our hospital, he'd probably actually still be on an oncology service. And so the main doctors caring for him would probably be oncologists, and there would still probably still be tons of doctors coming in and out who, you know, had all the attributes that you describe. On the other hand, there'd be a primary care doctor coming in every day. And talking to you and him and knowing him and able to help translate what all these other doctors were doing or redirect them to some degree if it seems like what they are doing doesn't make sense. And then if he gets--and then, when he's released from the hospital and goes to clinic, that doctor would see him in clinic and know what had happened in the hospital. And then the next time he came back to the hospital, that doctor would see him again. Whereas, it's very possible that you might be under the care of a perfectly[?] different oncologist on next hospitalization. So, those would be what I would say would be the minimum changes. Now, you can go further with this sort of model, and even eliminate more of the care of the specialization. You can imagine a specialist being in essence a Comprehensive Care Physician, seeing the patient in and out of the hospital. And there are medical specialties, I can tell you, in my institution, where they do that. They actually take care of the patient--same doctor--in and out of the hospital. So, there can be those sorts of models. But they are the exception rather than the rule. It's also true that doctors make lots of decisions about whether or not to consult. And there's a variety of drivers of that decision about whether or not to consult. But, let's say that, you know, a patient--imagine your dad--had rheumatologic condition. And there was some question as to whether that rheumatologic condition was contributing to some symptom he had or something like that. Well, if the doctor caring for him in the hospital is a doctor who doesn't really know him, he might way consult a rheumatologist. On the other hand, if the doctor caring for him is the same doctor who has been caring for him in clinic, he may well have seen these same symptoms before, have talked to a rheumatologist already in the past, know that these symptoms are the same old symptoms that have been seen in the past, and just avoid, you know, at least that part of the visit. So, would this model get rid of all of that fragmentation? Absolutely not. Nor should it. Division of labor is a good thing--
Russ Roberts: Adam Smith--
David Meltzer: to a degree. Yeah. Absolutely and this absolutely goes back to Adam Smith. But, it can go too far.
Russ Roberts: Yeah. There's something really beautiful about these specialists trooping through who know an immense amount about their one thing. But there was a point where I remember thinking to myself, 'Who is in charge here?' And of course, the answer was really me. Which is not a good idea. It's not a bad idea. But not everybody has a son visiting them who knows about tradeoffs. And risks. So, one of the things that struck me as strange about the whole experience is that--you know, a specialist would come in. The specialist might know something about the other parts of the body that were, you know, struggling in the aftermath of the surgery, and he might have the [?] to verify that. But each doctor is making their own decision about, 'Oh, yeah; you don't need to worry about x,' and I'm thinking, 'Well, wait a minute. I want to ask that other guy, the other person, about whether that's a good idea or not.' And, it just--there was no one in charge. And so one of the minimal benefits of this approach, beyond the knowledge of the patient's richer history and habits and pre-existing conditions is simply someone to help coordinate the decision-making in what is an n-variable, complex system of how to respond in the aftermath of a traumatic event.
David Meltzer: Yeah. And I will say, though, that in any hospital, there should at any point in time, be a primary physician--and by there I do not mean primary care but a main physician under whose care you are. And all the other doctors are consultants. So, on paper, to the extent it's not the patient or their family in charge, but to the extent there is a doctor in charge, it should always be pretty clear who that is. There are what are called co-management models, where sort of the attending of record, for example, might be a hospitalist, but the specialist is really in a lot of places the person running the show. But, when those models exist and they are even marginally functional, there should be very clear rules of communication between those people. So, um, you know, I think that a model like this should improve that coordination. But, to be clear: That coordination should always exist.
Russ Roberts: And I don't mean to suggest it didn't exist in this case. I think his surgeon was "in charge" in some paper sense. But, on the ground, in that room, at 2:30 in the afternoon when the GI [gastrointestinal] guy was there, it might be different. It doesn't always work as well in practice as it does on paper.
David Meltzer: Absolutely. And nothing, including the Comprehensive Care Physician model, is perfect.
Russ Roberts: Yeah, of course. So, this is very interesting; and of course there's a potential that mental health results could actually have health results beyond just it's nice that they were more satisfied with their doctor and with their general wellbeing. It's a short study so far. We don't know what the longer effects will be. But it sounds very promising.
David Meltzer: Absolutely.
Russ Roberts: So, congrats; and we look forward to the next chapter; and we look forward to that.
Russ Roberts: I want to shift gears; and I want you to put on your economist hat for a minute. My son needed an MRI [magnetic resonance imaging] recently. It's great when EconTalk is a family history of the Roberts family medical issues--sometimes it devolves into that. So, my son needed an MRI recently; and he's on my health care still; and we have a high deductible. And I said, 'Shop around.' So we called one place; and they said, '$8000.' I said, 'That seems like a lot.' I said, 'Call a second one.' And that second place said, 'It will be $3000-$4000.' I said, 'That still seems remarkably high.' I actually called a doctor, and it was a concierge doctor; I said, 'Where would you send him for an MRI?' And she gave me recommendation; and they said, 'Well, if you pay out of pocket, we'll charge $500.' So, that's startling. It's not a new finding. I'm not uncovering anything--everyone knows this, that pricing in health care is crazy. But, it's such a dramatic example of how the lack of price transparency, the lack of out-of-pocket payment, and the lack of competition, because of those things has created this crazy quilt of pricing. And presumably of quality, although in this case I suspect they were all pretty much the same. What are your thoughts, as an economist--and one trained at the U. of Chicago, where--I think I was there a little before you, but we were really big on competition. A lot of people don't like competition in medical care. It scares them. What are your thoughts, now having a Ph.D. in economics and being in the field of medicine?
David Meltzer: Well, I think what theory says is that competition in most instances, in the absence of things like insurance, is going to improve quality and lower costs. And by and large I agree with that. You can tell other stories in health care, where, when people are insured there's a tendency for organizations to kind of compete on quality and to promote greater utilization rather than less. So, you can tell other stories. But, my general belief is that more transparency is a good thing and that more competition is a good thing, too, in health care. Now, there are challenges. One challenge is the quality of the information. Another challenge is sort of the price system, which is, as you know well, not all that rational from a whole variety of viewpoints; and that makes competition particularly complicated. I also think it's really important to recognize the connectedness of so many different parts of health care. You know, a diagnostic service is not that easy to unbundle from a therapeutic service; quality is difficult to assess. Also, recognize that risk[?] is a meaningful experience, and that health care is the number one cause of bankruptcy in the United States. High deductible plans are a significant problem with respect to risk for patients of economic cost. They are also becoming an increasing concern for health care providers with people not being paid. You know, there are laws in the United States like, EMTALA [Emergency Medical Treatment & Labor Act] in particular, which says that if you go to the emergency room, they have to treat you if it's an emergency regardless of payment. I like living in a country where that's going to be the case, where people get care in an emergency whether they can pay for it or not. But, I also recognize that that's a challenge, economically. And, it's interesting. So, look: What's the ideal world? The ideal world from an economist's viewpoint is that we are all super-informed and understand exactly what we're buying at every moment or have people we pay to help us figure that out. That information flows costless. That we all have enough money in the bank to be able to handle risk. And so we bear the costs of the decisions that we make. You know, we are very far from almost every one of those things.
Russ Roberts: Of course.
David Meltzer: Let me just add one more thing. There's kind of--theorem of the second best, which says that if you've got a market inefficiency in the context of other market inefficiencies, there is no guarantee that eliminating one market inefficiency will improve welfare. And, our problem is that we have a whole pile of market failures here. And so, there is in my mind not a strong theoretical justification for simple market-oriented interventions in health care without their empirical analysis. In other words, you can theorize all you want about any uni-dimensional simple fix in health care that moves ones towards some idealized market model; but the reality is, the only way to know if it works, is to study it.
Russ Roberts: Yeah; I don't disagree with that. I'm more worked up with the fact that the default is always to move away from markets, it seems to me. I'll phrase it a different way. So, this is not as harsh as it sounds. So, you found something that has potential to have some significant savings. And my response to that is: Who cares? And, I don't mean that as a snarky question. I mean it as a statement of a problem that our health care system has, that: So, who is going to have the incentive to adopt? Let's say it's true, that a step in the direction you've hypothesized is a correct one. It's one experiment. We know that not all Randomized Control Trials generalize. But, it's what we've got.
David Meltzer: But let's say it turns out to be true--
Russ Roberts: So, like many other problems in the health care market, where--you know, I've recently had Vincent Rajkumar on the program talking about the price of multiple myeloma drugs and how a tiny increase in efficacy, like 'extends life on average two months,' leads to a remarkably large increase in expenditure because no one really has the incentive to capture that savings; and the gains are small relative to the cost that normally would never be done. But [?] they are done. Because nobody cares. There's no good[?] incentive in the system to take steps that are making it better.
David Meltzer: Yeah. So, I think--well, who benefits from this? First of all, I'd argue the patients who enroll in it benefit from it. Okay? So, that's a really good start.
Russ Roberts: Agreed.
David Meltzer: So, they would be interested. Now, they don't get that offered to them unless someone is willing to take these jobs; and people aren't going to take these jobs presumably unless the right conditions are made for them to do that. So, who is going to benefit? Well, I mean Medicare is going to spend less money. That presumably is attractive to Congress. And I would hope that the Executive Branch of our Federal government in general. So, I would hope that Medicare would see an intervention that seems promising, and want to try it out. And in fact, we've sent in a proposal to Medicare for what's called a physician-focused payment model, and they are reviewing that proposal now. And, we hope that within a couple of months they'll give it a formal public review and make a judgment about it. And if they decide it's sufficiently promising the way we've proposed it, actually adopt a payment model that they would test and that could conceivably produce savings for Medicare while also improving patient outcomes. Now, the nature of that model is to provide some very modest payments to doctors who reorganize their practices, to try to encourage this sort of model of care. But, some of those incentives are already existing within accountable care organizations or others. So, we kind of view this payment as sort of a nudge that pushes that ahead a little more and lets us test whether this works. And we think that if it's proven to work, then in fact those nudges may need to be, can be even smaller over time. So, you know: Are there people who might not like this?
Russ Roberts: Hospitalists are the first group that comes to mind.
David Meltzer: Well, actually, you know, it's interesting, because who is better suited to do this work than a hospitalist?
Russ Roberts: It could be.
David Meltzer: You know; and in fact most of the doctors who do this are hospitalists. And I'll tell you honestly: you might imagine that the hospitalist [?] the hospital, but that really has not been my experience. I think that even primary care doctors--I think there are some doctors who hear about this model and they feel like maybe the model is going to push them out of providing [?] in-patient care. But the answer really is, 'No'; in fact what we've been doing is trying to help primary care doctors who want to sustain in-patient practices to do this.
Russ Roberts: I think there's hope, David; don't misunderstand me. You happen to have stumbled on or designed--I don't know how much of it was serendipitous. You've found an improvement that could easily be adopted. And could--
David Meltzer: Yeah--
Russ Roberts: But in general--
David Meltzer: I agree. In general, there are very strong, powerful forces in health care that make it difficult to make change. I mean, even in this model there are challenges. We have a reimbursement model for risk that doesn't do probably a particularly good job of reimbursing for the sickest patients. There are structural challenges that one has to deal with. And, we've tried to design a payment model that addresses some of those things. There's inertia. There's fear of change. There certainly could be hospitals that look and worry it's going to [?] decrease the number of patients in their beds and [?] decrease--
Russ Roberts: Yeah, they're not going to make their numbers and then their investors [?]--
David Meltzer: Yeah. But then other hospitals are already taking risk and recognize that their long-term value to payers whether private or public, and in patients, is to improving health and increasing value. So, you know, there's the short run and the long run. I think in the long run the alignment is pretty good. I think in the short run there are things to be proven and tested and figured out. But, you are right. There are tremendous entrenched interests in health care, and so interventions that produce social value have to not only produce social value but figure out how to overcome the potential resistance of range of parties that have the ability to make that intervention not be implemented, even when, on net, there is benefit.
Russ Roberts: Extremely well said. Very eloquent. Beautiful.
Russ Roberts: Well, let's close with a, if I might, a personal question. I'm ashamed to say this: I can't say this authoritatively, and I should be able to, but I think you're the first M.D./Ph.D. that I've had as a guest. I've had a lot of economics Ph.Ds, obviously--a lot of economists. And I've had a reasonable number of doctors on the program. But I think you're the first one--I apologize to any guests in the past that I've missed or not thinking of. Why did you do that? It's an incredibly expensive investment, obviously. A great deal of fun. We both had Gary Becker as an advisor--that's just an interesting human experience that I'm grateful to have had.
David Meltzer: Me, too.
Russ Roberts: But it's incredibly expensive. And, I'm just curious if you could tell us why you did it, and how, in your day-to-day practice as a physician, how being an economist helps you or hurts you. What's it like? Because it's unusual.
David Meltzer: Yeah. So, my story about this was that I was a Yale undergraduate. I double-majored in economics and in molecular biophysics and biochemistry.
Russ Roberts: Perfect!
David Meltzer: And, I'd grown up in Chicago, on the South Side. I was incredibly interested in economics and social science and public policy. But I also really liked science. My dad is a biological psychiatrist studying schizophrenia. So, I started in medical school--I'm sorry--I started college thinking I might be interested in chemistry or biology, but also social studies. I took economics; I found I really loved economics. I took science. What I learned was that I loved the rigor of science, of basic science; but I actually didn't care how the experiments turned out. In contrast, I found economics just constantly stimulating. But, at the time in particular, economics was incredibly theoretical. And it felt like a career in economics was going to be all about theory and not about applying it to real problems. So, I knew about MD-PhD programs. And, at the time, I had thought about sort of using the science. My dad gave me the advice that if you're interested in science, biomedical science, you should be a doctor rather than a Ph.D. because Ph.Ds have to work for doctors, and doctors are terrible to work for. So, he recommended that I think about MD-PhD programs. I suspect also knew well that MD-PhD programs were pretty well subsidized by Federal government. But, I wondered to myself: Could you do MD-PhD in economics? And--
Russ Roberts: Because the more standard one is in one of the sciences--
David Meltzer: Is in the basic sciences. So anyway, I wrote letters all around the United States. A third of the places said that's a really stupid idea. A third of the places said, 'You've got good grades: if you want to do a real Ph.D. in the sciences please apply to our program.' And a third wrote back and said, 'Strange idea, but reasonable. Please apply.' And the U. of Chicago wrote back and said, 'We have a program.' And so, you know, I later discovered no one had ever actually been in it. Anyway, I had a wonderful economics teacher at Yale, Paul Schultz, who was the son of Ted Schultz, another U. of Chicago economist. And Paul taught me this amazing course in economic demography and from that I learned all about sort of Chicago economists studying population and health and Gary Becker's stuff. And the whole crew. And so, coming back to Chicago was sort of a great thing, because there were just so many wonderful economists--in microeconomics, labor economics, industrial organization. And I got into medical school and the Ph.D. program; and it was coming home. So, I ended up coming to Chicago. And then Gary ended up being my Thesis Chair, along with other members of the committee--Bob Willis, Sherwin Rosen. You know, great folks. An amazing group of people around Chicago in that era. I don't even know how many Nobel Laureates taught me economics. It was a wonderful exposure. And it's interesting, because when I look at the Comprehensive Care Physician work, it is literally the tools that I learned in Gary Becker's microeconomics class that I'd use when I sat down and tried to figure out the time allocation of doctors. And then that gave me, literally looking at those equations, the idea for the Comprehensive Care Physician program. So, I give an immense amount of credit to them and their way of thinking for inspiring the work that I've done. Of course, it's had a very practical, hands-on component to it, too. But I think that the sort of nugget or the insight really has come from my training as an economist. You know, we have an MD-PhD program in the Social Sciences at Chicago now; and we have had some amazing graduates who have done the MD-PhD like I have, and gone on to do incredible work. And so, there's now quite a few more MD-PhDs in economics around the country and around the world; and I think they are making a lot of great contributions.
Russ Roberts: But in the day-to-day practice as a clinician, are there things like the kind of pricing structure or incentives that you see that drive you crazy, because you understand their impact?
David Meltzer: You know, I mean, there are lots of things in medical practice, in academic medicine, that you can recognize as irrational. And I guess you could allow those to drive you crazy. But then there are also, on the better days, like really interesting ways of thinking about problems that economic reasoning teaches you. Like trading off competing values. Organ function as a durable good, as an asset that can be used and transferred to other forms of assets. Like, you know, when you have a patient in the intensive care unit with multi-system organ failure, they can have a lot of renal capital but not very much cardiac capital. Right? And so you can do things to protect their heart, at the expense of their kidneys, with the idea that, you know, that sort of scarce cardiac capital is more important. So, economics is such a powerful way of maximizing objectives subject to constraints. And I think that's an example of what you were describing earlier, about how the body and the economy have certain similarities. And the tools for thinking about each are complementary.