 Good afternoon, it's a pleasure to welcome you to this first lecture in the lecture series this year on the present and future of the doctor-patient relationship. On behalf of the McLean Center for Clinical Medical Ethics, the Center for Health and Social Sciences that is chess, and the Bucksbaum Institute for Clinical Excellence, David Meltzer and I welcome you to this program. The lecture series will consist of 26 lectures and will meet during the fall, winter and spring quarters. In an era of advancing medical technology, the series will examine whether the doctor-patient relationship is likely to survive the challenges of the 21st century and if physicians will be able to continue to practice personal and caring medicine. Speakers in our series will include Dr. Roman DeSantis, a great cardiologist from Harvard who practiced there for 60 years and has written a book on doctor-patient care, Professor John Harley Warner, the chair of the history of medicine at Yale, and Dr. Holly Humphrey, president of the Josiah Macy Foundation and a former dean of our medical school here, among many others from outside of our community. I'm very excited, as is David, about this year's series, not only because of the quality of the speakers we've recruited, but also because we have been in contact already with Oxford University Press and Springer, and even with the University of Chicago Press, talking about the possibility of publishing the outcome of this program. It's my pleasure now to introduce today's speaker, our keynote speaker in the series, David Meltzer, who, as I say, worked very closely with me to work out this and arrange the series. David and I go back a couple of years. I measured about 25 when he was a fabulous medical student here at Pritzker, and when at Pritzker he also did his PhD in economics, working with the late Gary Becker. Dr. Meltzer is the Fannie L. Pritzker professor of medicine, the chief of the section of hospital medicine, and director of the center for health and social services. David's research explores problems in health economics and public policy, and recently his research has focused closely on the doctor-patient relationship. David also serves as faculty in the Chicago Harris School of Public Policy and the Department of Economics. Currently, David is a recipient of a Center for Medicaid and Medicare Innovation Challenge Award and is leading an effort to study the effects of improved continuity of care on the doctor-patient relationship, and the study relates to the same physician caring for frequently hospitalized Medicare patients, both in the inpatient and also in the outpatient setting. So it's a huge pleasure to welcome Dr. Meltzer. The talk he's giving today is entitled Nuage, the Economics and Ethics of the Doctor-patient Relationship. Please join me in giving a warm welcome. Thank you. Thanks, Mark. I think it's actually more than 25 years. It may be. I'm too, my brain is atrophied too much to do the math, but it's a lot. So I'm really excited for today's talk and for this series. I have been aware of the importance of the doctor-patient relationship and interested in it since I met Mark as he says at least a quarter century ago. And I've been doing this work that he referred to on the the the doctor-patient relationship through this comprehensive care program. But today I'm going to sort of frame this in a different way. And I'm sorry, let me, there's a something that's cutting off the screen here. Is this a touch? There we go. Okay, now I can see the screen like you can. So the title of my talk today is Nuage, the Ethics and Economics of the Doctor-patient Relationship. And let me start by telling you what I mean by Nuage. And the first thing is I don't mean nudge. So what is nudge? Nudge is a word we've heard a lot in recent years, particularly perhaps around the University of Chicago, but all around the world really. And it's a small action, such as a change in a default or a reminder that results in some sort of change in behavior. And it was popularized by this book by T University Chicago faculty, Richard Thaler, who's an economist and psychologist who has sort of popularized the field in many ways of behavioral economics, and Cass Unstein, a law school professor who was here now at Harvard. It had an incredible impact in bringing this idea into the public space. Now, I will point out that they didn't come up with the idea of behavioral economics, although in fact, Thaler had been making fundamental contributions for a long time. But they sort of mobilized the literature. And I want to give credit to someone, almost none of you probably know, Bridget Madrian, who was a professor here in the business school when I was a junior faculty member. She's now the dean of the business school at BYU. And she did a wonderful, very simple study that I think really catapulted forward this idea of nudges. And this was, she basically experimented with changing the savings default in people's paycheck, so that if you did nothing, you were automatically opted into saving into a retirement plan, and you could choose to opt out. As opposed to having the default be just you got your paycheck and you had to actively decide to save. Now, economic theory says that little default shouldn't matter at all. She showed it mattered immensely. And that set off an entire literature on this. And then that was fueled by technological changes, an iPhone shown here, but iPhones, computers, all sorts of technology that make it easier and easier to change defaults, to remind people of thinking, fit bits, a million other little tiny things that can drive potentially behavior. Now, as we talk through this literature, I'm going to spend the beginning part of the talk talking about nudges. One of the key things that I will argue about this is the nature of this is that it is a very small push, not a big effort. And so it can easily be done on a broad scale. I will argue it is also not very effective. And so hence it is on that not very important. I want to contrast nudge with nudge. Now, I don't know if everyone in the room knows what a nudge is. It's a person who pesters and annoys you with persistent complaining. It's a Yiddish word. It comes from, as you can see here, to pastor. And it has background in Polish and Russian. And it's to wear out with complaints or questions. There's also a similar word in Old English. Given that it's a Yiddish word, you can't help but thinking the Jewish mother or the Jewish grandmother. I have one of those. I had at one point all of those, at this point I only have a Jewish mother. But I don't have this type of Jewish mother. This is a picture from a Woody Allen movie. That's the back of his head if you don't recognize it. And the movie is called Oedipus Rex. It's part of a series of three movies called New York Stories. And in it, his mother disappears suddenly and appears floating over New York City visible to everyone, talking about her son. In particular, why won't he marry a nice Jewish girl? And she never disappears until finally he does. That's a nudge. It's this thing that's always there. And I will argue that the doctor-patient relationship, although I don't mean it just imply that doctors are annoying or should be annoying or should pester, what I mean to emphasize is that whereas nudges are these little things that go away, that relationships can be big things and stay. And I will argue that big things that stay in the end are far more important than little things that go away. And I think that's the key message. And I think it sets up in many ways why this series on the doctor-patient relationship and what it is and why it works and what are its threats is such an important one. Now, I'll come back to the figure floating above in a minute or maybe at the end and we'll talk a little bit about what that really means. But let me first turn to nudging. So as I mentioned, this literature on nudging has completely exploded over the past few years. And there was within the economics literature just this year a really comprehensive review of all the randomized trials done of nudging to date. Turns out 317 different randomized trials. And what the authors do is break it down on the rows there and sort of the category of nudge. So like changing a default, that was the one I mentioned that Bridget Maydrian did, simplifying things, getting rid of too many choices. Some of them are sort of weird like changing effort, making it easier to do something. That kind of is a nudge in the sense of being a small change often. But I would argue that rather than thinking of that as sort of behavioral economics, I would just think of it as economics, that time is money. And so making things cheaper or easier isn't necessarily the same as a nudge, but that may be a distinction that's not that important to you. For those of us who are trained as economists to call prices a nudge or making things easier a nudge doesn't really meet behavioral economics. Warnings, disclosures, so it's largely about information. Pre-commitments, really interesting idea, you put money up front and then you don't get it back if you don't do something. All these sorts of things. And you can see these different approaches have been applied in a variety of different fields, health, environment, energy, privacy, finances, policymaking and a bunch of others. There are also some patterns I may be able to come back to a little bit in which of these approaches have been used in different domains. So this is some figures that show the growth over time in these studies by category and of nudge and the context in which they've been applied. And what you can see is, in the sort of 2010 era, not a whole lot of randomized trials coming out. And then in the past four or five years or so, just a huge explosion. This is in the left panel. And when you look at them, what you'll see is in the past few years, defaults have grown quite a lot. That's the blue bar at the bottom. Also warnings and graphics, really just provision of information. I'm not quite sure that's really behavioral economics, but defaults are clearly a very important part of this. The other thing that you'll see if you look at the growth by context over time is you can see that health is, that's the blue bar at the bottom, has been fairly big. It grew around 2014. It was big early on, sort of grew about 2014. And that's actually been fairly flat. And other applications, environmental policy, finance have grown more. So people are interested in this, but yet it's not growing as fast as the field as a whole is, which I find interesting. So how effective is nudging? So what they did in this meta-analysis is they looked at the distribution of the effects of these across different categories. So the heavy bar in the middle of the box and whiskers diagram is the mean. And you can see that in some areas like finance, the line's pretty high. Actually, in environment, it's pretty high. Privacy is also exactly what the privacy category is. But finance and environment, it seems to have a pretty big effect. Energy policy, not quite so much. Health is kind of somewhere in the middle. It's not the most. It's not the least. What about different, oh, and let me just point something out. The mean effect size is about 20%. Okay? Now that's a peak effect size. I'm going to come back in a moment to how long it lasts. These are the different approaches. And one thing that's interesting here is that some of them, the clear winner in this is defaults. So where Bridget's work began is still the thing where you see the largest effect. That's the second bar from the left. Okay? So setting the default seems to really be something that seems to work on a fairly large scale. I'll also just go back and point out default is not super prevalent among health. It's up there, but it's not that big. So the most effective thing that seems to work in these isn't used all that much when it comes to health. Now, why is it? I would argue that it gets back to simple principles of ethics. Default in a way, takeaway autonomy. It raises questions about all the other elements that we worry about as core elements of ethical medical decision-making. But clearly autonomy is something that is at risk when we change defaults, particularly if we don't follow them with active engagement. And I think the whole point of a lot of the medical system, in fact, the medical legal precedent in a whole variety of ways is the need to, in fact, engage patients and activate their autonomy. And how much defaults really work in that context, not clear at all. The other big problem with nudges is they don't work for very long. So this is another analysis of these sort of RCTs and their effects. And they look in a variety of different areas, charitable giving, education, exercise, smoking cessation, weight loss, and so on. And what you see is the x-axis is weak since the end of the intervention. And effectively, the one is the initial effect size. So think 20 percent change in the behavior. And then all these things just collapse within weeks. So they just don't work long. And that's a big problem. I think the problem in health though is even bigger, which is that all of these studies about nudges target very specific things. And is health a specific thing? Well, the World Health Organization defines health quite simply and narrowly as the state of complete physical, mental, and social well-being and not merely the absence of disease or infirmity. So in other words, you have to change not one thing. You have to change a whole bunch of things if you're really going to make someone healthy. Now, you can say that's a lofty WHO definition. But look at concepts of patient-level centered care. Read the definition. In patient-centered care, an individual's specific health needs and desired health outcomes are the driving force behind all health care decisions and quality measurements. Patients or partners with their health care providers and providers treat patients not only from a clinical specter, but from an emotional, mental, spiritual, social, and financial perspective. So the idea that you can find a very narrow nudge, have a computer program push that little information in that personalized way to that person at that moment really does not seem likely to me. And so the goals of health care are extremely complex. Now, why does complexity of goals matter? There's this an economist, John Tinbergen, who's a Dutch economist and interestingly was the first winner of the Nobel Prize in economics. He had some very interesting interactions with the University of Chicago School of Economics. Several of his students were faculty members here. And Tinbergen brought this very strong belief in policy and social engineering. And then his students came to the University of Chicago and of course encountered free market economics. What Tinbergen is most known for is what's called the Tinbergen rule. And the idea is that if you have multiple policy objectives, you need multiple policy tools. And so he believed that you needed to have multiple interventions and you need to understand how each one influenced particular things and then combine them together. So now you put together what I said about health. It's this multi-dimensional complex thing. And then imagine what combination of iPhone apps is going to get you to that. Okay, pretty challenging question. Okay, so are there other simple nudges or pushes? And I wouldn't call these nudges. I would call them incentives on that change health care. And the leading one in the United States is the belief in prices that if only we fix the health care system and its economics, we're going to solve our health problems. So what do we have to teach us about that? Well, we have a lot of empirical studies, but one of the most notable is the Rand Health Insurance experiment. The results of that are published in a book for any of you who care to take a look at it. Joe Newhouse was the leading author of that. That's Joe's picture nearing the end of the study in the far lower right. He gave a talk here just a couple of months ago. He's still on the faculty at Harvard. Above him on the left is Nyua Duan, who's a very good friend of Robert Gibbons for any of you who know Robert. Ask him about Nyua. And does anyone know the top right? It is Will Manning. Thank you, Greg, for calling it out. That is a young Will Manning who was on the faculty here until a few years ago, and a lovely man who we truly, truly miss has passed away. So what did the Rand Health Insurance experiment do? They randomized a bunch of adults to different levels of co-payments, and they saw how that affected their health utilization and outcomes. And what they found was that higher co-payments decreased utilization, but that they had potentially important effects on health, particularly things like controlling blood pressure, which turned out to really matter in the effect of health care on health outcomes, and that these reductions were biggest among low income people. So this is the Tinbergen problem again, right? You have a tool, but you have a problem that goes right along with that tool. There was a wish, I don't hope, that if you better covered outpatient and preventive care, you would actually save money, that you would prevent people from getting ill, keep them out of the hospital. The Rand Health Insurance experiment did not show that. That was the most striking disappointment. So that tool, again, does not work particularly well. Now there was also a branch in the Rand Health Insurance experiment where they looked at managed care. And what they found was that managed care did pretty well in reducing costs, but it was associated with worse health outcomes, particularly for low income adults who were in poor health. And that sort of study result is prevalent in many, many, many places in the literature. And I would argue that even Kate Baker's more recent work in the Oregon Health Insurance experiment doesn't really refute that in a lot of ways. Certainly she doesn't show improvements in health outcomes with expansion. But it's also very clear that there are not benefits in terms of reductions with this access. So there is, even more interestingly, some recent work looking at the effects of expansions of coverage, insurance coverage in the Affordable Care Act on health outcomes. Now Kate's work is largely, widely interpreted as saying Medicaid doesn't really make a difference. Okay, that actually, at least it doesn't make a difference in terms of health outcomes. What's interesting about that study is it has very few really vulnerable people in the study. And what's interesting in this new work, and Laura Wary is the person who has one of the lead authors of this, and Laura was a PhD student in the Harris School. What they're showing using census data is that among older adults, so people who are just below Medicare age, so 55 to 64, there actually been substantial reductions in mortality with the passage of the ACA. So watch for this. I think it's really important. So the sort of expansion of coverage really does have benefits. But again, with managed care, you may not get these benefits from people. So I take all those things together, and I conclude that nudges, whether they're behavioral economics nudges or small economic incentives, are not going to get us where we want to go. So what about the nudge? Well, again, I think it's not pestering people, but this continuing relationship that matters. And so what I'm going to do now is try to give you a little bit of an update on the work we've been doing in the Comprehensive Care Program and show you some of the things that we've learned. And there's a variety of other things that will come out of this. And I also want to mention that in case I forget at the end that Joyce Tang has been doing some wonderful work looking at the nature of the relationship. And I'm going to let her talk about that when she's ready. But I'm going to argue that the building of these relationships through that program is producing really noticeable benefits. And then hopefully Joyce and others will help us figure out why. So where does the sort of program begin? It begins with this sort of understanding that there is a challenge of sort of the care of very complex patients, their fragmented care, and the failures of care coordination. And this begins with the idea that there's a very small fraction of patients who account for a large fraction of health care spending and adverse outcomes. These patients often have multiple interacting conditions that involve multiple specialists in multiple treatment sites, ultimately fragmenting care. And a large fraction of these costs and adverse outcomes are tied to hospitalization. So improving inpatient and outpatient care coordination seems to be a key opportunity, but it's turned out to be really challenging. A figure on the left is some figures from a congressional budget office report on Medicare's experience with care coordination programs. We are all familiar with these programs. You typically have someone, a care coordinator, a social worker whose job is to communicate among the different doctors in the team and connect with the patient. And the unfortunate reality of these programs is that although some of them have decreased hospitalization rates, just as many have been associated with increases in hospitalization rates, and the congressional budget office has sort of concluded on average they don't consistently work and don't save money. And you can stratify them by how much interaction there is between the doctors and the patients, between the inpatient and outpatient doctors. In general, the more interaction, the better, but also the more interaction, the more costly, the less likely the money gets made back. So the figure on the bottom right sort of is my way of thinking about this. Think about the bars as money. We spend a lot of money on hospital care, a little money on ambulatory care. We want to shrink those bars. We're not going to shrink ambulatory care because that keeps people out of the hospital. We now need to pay for care coordination, so we have to make room for that. So we have to shrink the hospital care a lot if we're going to save money. And in the end, that's just not practical. And that's, I think, why these programs don't work. So that's kind of the challenge. And the challenge, unfortunately, isn't getting better. It's getting worse in a lot of ways. And unfortunately, the growth of hospitalists are making this problem of inpatient and outpatient care more difficult. This is a change from the traditional model where people at a doctor who cared for them in and out of the hospital, this sort of ongoing doctor-patient relationship, when it started, it was hoped to improve care and lower costs. People highlighted the advantages, such as inpatient expertise and presence of these hospitalists. But they also noted the disadvantages, distant continuities of care, loss of the doctor-patient relationship, and I'll come back to that in particular in a moment. And when you look at the literature, and we've contributed to this literature, the net effect seems to be modest. So if hospitalists aren't really much better, why did they grow so quickly? And one reason perhaps they grew is just people believe they were better and they just weren't. But in recent years, we've been doing work emphasizing another theory of their growth, that hospitalists, instead of growing to meet the needs of hospital care, grew to meet the needs of primary care. That with decreasing hospital volumes over time relative to ambulatory volumes, which was mostly because ambulatory volumes rose, it was harder for doctors to block out their day to their morning to be able to see patients in the hospital. Their clinic was fuller and fuller with lower and lower acuity patients. They could see patients all day long in clinic if they wanted and yet have very few patients in the hospital. So it made sense for them to kind of turn this over to some colleague who eventually became the hospitalist. And so this figure on the bottom right is from a paper we wrote showing over the course of the 80s and into the early 2000s an 80% decline in hospital volume compared to ambulatory volume for the typical ambulatory physician. So that's kind of the challenge. Now what I've just described is a theory. We've tested this theory in a rigorous way with a sort of ambulatory economics model of hospitalist growth. In this we compare the time costs of two models. In one there's a traditional model where internists spend their time seeing patients in the hospital and in clinic and then have to drive in between the two so they have transport costs. In the other new new hospitalist PCP model the hospitalist spends their time seeing patients in the hospital but now has to communicate with the PCP and the PCP spends their time in the clinic but now has to communicate with the hospitalist. So the difference in cost between the two models is driven by per capita communication costs relative to transport costs for a typical internist. And all the math here just simply formalizes that. And that then proceeds to give us a series of testable hypotheses about which primary care doctors will turn over their work to hospitalists. In particular the theory would predict that those who have fewer admissions relative to ambulatory visits will turn over to hospitalists. As communication costs decline it's easier to just call the hospitalist transport costs rise. Think parking in our parking garage and as physicians work fewer hours. These are all pictures from the Marcus Welby TV show of those forces actually playing out over time him seeing a healthy patient on the phone driving in the car and then the young doctor who took over him and like to ride motorcycles and not work long hours. And so all these trends were playing out and when we tested this with data every single theoretical prediction of the model was supported. Okay so I believe that hospitalists grew not because hospitalists were better for patients but because it was more convenient fundamentally for doctors. Okay so what's lost and what's lost I think is the doctor-patient relationship and you know why does that matter. Well you're going to hear about that you know throughout the rest of the year but very briefly there is a rich literature on the value of the doctor-patient relationship it emphasizes their trust the interpersonal relationship between the doctor and patient their communication and the knowledge of the patient. We also know patients value seeing their own doctor in the hospital they won't always pay a lot for it but they value it. There are also some extraordinary observational studies so Medicare patients who are cared for by their own PCP for more than 10 years have 15 percent lower costs lung cancer patients cared for by their own doctor in their terminal hospitalization have 25 percent lower odds of ICU use which is presumably feudal care so it's kind of better end of life care. These are observational studies you can criticize them based on that. My favorite study in this area is an experimental study by John Wasson in the 1980s where he randomized 800 complex patients see the same physician versus different physicians in each primary care visit and the continuous care group had this extraordinary set of changes 49 percent lower emergent hospitalizations 38 percent lower hospital days 74 percent lower ICU days. So I conclude from this that discontinuities harmful and costly especially for these frequently hospitalized patients and wondered is there a practical way to restore continuity in the doctor-patient relationship and this is where the idea of the CCP model came in. The idea is we stratify patients according to expected hospital use those with low expected hospital use get an ambulatory based primary care doctor and the rarents they're hospitalized they get a hospitalist but that doesn't happen very often so we don't need to worry about it very much in contrast those with high expected hospital use gets this comprehensive care doctor who can spend their morning seeing patients in the hospital but limits their ambulatory practice the patients at high risk of hospitalization so they can see those patients in the afternoon and give them all the primary care they need with a much smaller panel. What are the advantages the most frequently hospitalized patients get their own doctor in both settings patients value this continuity it decreases unneeded testing and treatment and errors it's much easier for the doctor they can just walk in and ask the patient what happened yesterday rather than spending their time trying to figure out their whole history. All hospitalized patients get doctors with significant hospital experience and presence these can even be specialists patients can choose their doctor which they can't with a hospitalist this can work practically for a physician with a reduced panel size you can use this with a patient-centered medical home or bundling or readmission penalties that sort of create incentives for savings and you can even build a small primary care base for a hospital that needs to fill beds and help it stay in business so what are the challenges will doctor will patients switch will doctors let patients switch will doctors do this job and can be economically viable to begin to answer some of these things we began in 2012 CMOI funded 2000 person RCT of the CCP program at the UFC these were some of the key elements of it this focus on patients at increased risk of hospitalization which we measured very simply with have you been hospitalized in the past year the small panel size of about 200 patients per doctor that allowed them to direct to interact directly with their patients seeing them on the wards every morning and seeing them in clinic in the afternoon a very tight and lean interdisciplinary team of five doctors working together with nurse social worker and so on this small these small well-connected teams to keep costs low some simple focus on health care transitions we did this original under fee for service then moved to financial more complex financial models like risk-based models we try to support the team for its very psychosocial complex patient population used rapid cycle innovation techniques and as I said rigorous evaluation with an RCT I'm just going to quickly remind some of you who've seen these before of some of the results we did this 2000 person RCT everything was sort of beautifully matched in terms of baseline characteristics between the CCP and standard care patients and then what happened the physician ratings so at the beginning of the study both the intervention and control group were not statistically significant from each other and they both started at about the 20th percentile nationally turns out that the control group goes up from about the 20th percentile of satisfaction of their doctor to about the 80th and and that turns out to be a significant decrease increase and that's simply because they didn't like their doctors before we got them a new doctor of the u.s.c for those of you who are doctors at u.s.c they like you okay um what happened in CCP it goes actually not from to the 80th to the 95th percentile and a statistically significant this shows out to one year it's actually stays the same thing out to two years so extraordinary increases in patient experience we didn't see differences in general health ratings between the two groups but when we looked at mental health ratings no statistically significant base at base difference at baseline and then as we go through the quarterly surveys we see um highly statistically significant improvements in self-rated mental health status but in terms of the sustainability of the model and its underlying economic implications the key thing is hospitalizations and here again no statistically significant difference at baseline between CCP um and control um but um once you get it to three months and beyond CCP is um statistically significantly below turns out to be about a 15 to 20 percent reduction in hospitalization rate sustained here out to a year and the effects seem to go out even longer so we concluded from this that it was possible to actually implement this program and patients like it it seemed to work okay for the economics of the hospital and that the care improved patient experience at least maintain their outcomes or reducing hospitalizations up to about 20 percent per year now this contrast with all these care coordination programs that really haven't worked now the number needed to treat here is pretty striking enroll four patients for one year to prevent one hospitalization so we estimate our program prevented about 250 hospitalizations in the first year the patients were enrolled in the intervention with a cost of about $15,000 per hospitalization that translates to about four thousand dollars per patient per year and this is actually a typo it's not one million dollars per year savings it's actually four million dollars per year savings for the program now there were limitations um self-reported outcomes although we had very high follow-up rate we've only had partial medicare claims data we're waiting for some more the medicare claims data is going to be complicated because there was some dropout of the program related to um Illinois medicare medicate alignment initiative but um we don't think that's going to change things and then most importantly it was this is only one ccp program it's one hospital a limited set of doctors and also a very unique socioeconomically disadvantaged population and that's where i'll talk about our newest program which i don't think i've talked about here before which is our comprehensive care community and culture program and the idea here was that we found that about 30 percent of patients who said they were interested in our study and actually signed up for it never enrolled or never really engaged they didn't make appointments they didn't keep them if we um if they did make them and so support from the robert wood johnson foundation we designed this comprehensive care community and culture program with three added interventions to try to better engage patients first of all systematic screening for unmet social needs with an instrument developed originally by health leads and 17 different domains secondly access to a community health worker to engage patients and connect them to resources and then finally access to community based arts and social program to try to activate the patient and the idea and get them involved in their own care and the idea ultimately was to line all these resources across sectors by identifying and meeting patient needs to activate them to engage in their care now here's our conceptual model for this we start with a high risk patient who may or may not be activated if they're activated to engage in their care and they're given access to ccp this is the top wrench they seem to do better and get improved outcomes but if they're not activated they don't really participate in the program so they don't benefit from it so we add these new elements in c4p the screening for unmet needs the community health worker in the artful living program to activate that patient get them involved in their care if this theory is right then the least activated patients should benefit most from the addition of these added services okay so we did a pilot study we did it for two reasons first to gain some experience operating the program but also to understand about unmet social needs and then secondly to do a pilot assessment comparing c4p versus ccp versus standard care with a bunch of different outcomes and we viewed it as being underpowered to look at outcomes but a chance to get some experience running the rct so what did we learn from the operational lessons I think most importantly that iterative development of the program elements was critical took us a while to recruit and train the team figuring out everyone's roles and responsibilities workflows and tools was critical but engaging the patients in the program was also critical for refining it this iterative experimentation you can imagine what a patient likes but unless you really talk to them and engage them you don't know and so that was key and what that means is it takes a while to get enough patients in the program to really figure out how to make it work we also over time learned that even our standard care needed to change so we became an accountable care organization here at u of c and so the comparison group changed over time we learned that it was impossible to engage patients about 30 percent of the patients actually did get involved the arts and social programming and there were a number of regulars but it was really interesting because the people who got engaged were often the people who were relatively less socially connected and then about two-thirds of the patients actually took advantage of the community health worker services and the social work data we also learned really important lessons from the unmet social need data as i mentioned we used a tool that looked at 17 different domains and you can see them here food housing money employment public benefits childcare all the way down to the list the original tool actually had only 16 elements we added the 17 which was spiritual and religious support because our patients told us this mattered to them one of the fascinating findings is even in this incredibly socially vulnerable population 90 percent african african american median family income in the mid 20s thousand dollars a year 59 percent of respondents said they had only zero to two unmet needs accounting for only 13 percent of unmet needs the other 41 percent of respondents accounted for 87 percent of the unmet needs and 24 percent of them had five or more needs accounting for 67 percent of unmet needs and when you talk to patients they said you can't possibly give me a half a dozen referrals for my half a dozen different problems there's no way for me to interact with all those organizations so we sought to try to understand could we develop interventions that address clusters of needs so we did something called a latent class analysis with help from students from the Harris school and basically we had i'm going to start from all the way on the right and go to the left we had a cluster of people who had very few needs we had a cluster of people a small cluster but interesting with like issue legal issues and child related issues fascinating in a medicare population a lot of these were grandchildren related issues we had a cluster not a huge cluster but a fairly needy cluster of issues related to money then second from the the left a need a cluster with sort of enjoying and engaging and enjoyable activities and healthy living and then finally in the first column only eight percent of participants but 30 percent of all these sort of everything's wrong cluster a whole set of interacting things and so these had a lot of implications for our program design and i'll come back to that in a minute so let me just show you the results of this pilot study and as i said it was a three armed pilot now our original study at about a thousand patients per arm this one was a little less than 200 patients per arm again beautifully matched in baseline variables no significant differences in demographic covariates or outcomes prior to the study enrollment and what did we see in terms of hospitalization rates well whereas previously we'd seen a 15 to 20 percent reduction in hospitalization rates now the odds ratios were like 0.73 0.71 so about a 30 percent reduction in hospitalization rates substantially larger down at the bottom you can see big increases in activation fascinatingly the activation was increased to both in ccp and in c4p so we think what's actually activating them is engagement with their doctor interestingly so that was an interesting surprise now we had a bunch of other sort of theories and questions like for example were the benefits all concentrated in those with the highest hospitalization rates it turns out no that even in the lower group of hospitalization rates so people in in our say in the study at least they were stopping benefits so that said we weren't too focused and in fact we've expanded the eligibility criteria in terms of hospitalization a bit in terms of theory remember I said this idea that we thought the patients who would benefit most from c4p are those who are least activated and that is exactly what we found the benefits were highly concentrated in that group the p values aren't significant here because we've now cut our sample size of 200 patients per arm down to 100 patients in each half of the arm but the trend here is very very compelling so conclusions from this that you know we were able to iteratively improve these operations of the program engage these patients even in the small pilot we found evidence of reduced hospitalization with c4p versus ccp in standard care we do need more sample size we do need to link to the claims data we also as I said found the sort of larger benefits of c4p versus ccp in the patients who are less activated at baseline it was notable that the activation improved comparably in ccp and c4p so as I said earlier this doctor-patient relationship the ccp intervention itself seems to be the activator here it is interesting that c4p did do better and we think that it's the community health workers that are making probably the bigger difference than the artful living program it's hard to know for sure and it's just because the rates of enrollment participation in the artful living program are a lot lower and as I said we saw comparable benefits of these programs in people across hospitalization rates so we're now expanding it to somewhat lower risk populations so next steps we're completing analysis of these programs here at U of C we're doing longer term follow-up looking at costs duels and non-duels on hospitalization rate doing qualitative assessments of the doctor-patient relationship Joyce is doing that as I mentioned Alex Tate's leading work on end of life care in ccp for the c4p data we're completing our analysis of the unmet social needs data including overtime and looking at outcomes data we're now launching a phase two program funded by rwj which will be a thousand patients per arm again of c4p versus ccp versus now a care coordination model in this we've tried to simplify the surveys to make it a little easier for patients to do long-term follow-up and we've also critically tailored the newest iteration of this c4p program to these clusters of unmet needs and you can see here remember before I described there were five clusters including a sort of basically no problems cluster we don't worry about them but again from right to left in this case we have this child-related needs group where we have a social worker who will focus on that a health and insurance and financial group and Harold Pollock and others are helping us develop a sort of financial literacy and intervention to help patients you know manage their finances better for healthy living and social engagement at mcdonald and other folks are helping us develop a healthy living program with cooking and community-based gardens and then finally for the sort of many basic needs group with multiple problems we've developed a social service alignment learning collaborative and my colleague Emily Parris is really leading that and in that what we're doing is we are bringing together social service organizations that have capabilities and literally all these 17 domains and bringing them into the room together i'm initially weekly and now once a month to talk in a case-based format about how we can together better care for these patients who have these multiple needs and some of it is about organizations working together but I think a key message from this is that these patients who have these multiple needs need a person they need someone who can help them navigate across these organizations and very few of these organizations are incentivized or resource to provide that sort of deep personal navigation and so that's a really interesting question and again is very much analogous to the nudge versus nudge right it's the continuing relationship with a provider in terms of dissemination efforts for these programs we're busy here at the University of Chicago trying to migrate this from fee for service into value-based contracts we're busy expanding this out to Ingalls Hospital which i'm really excited about as an opportunity to do this there's been a lot of interest in this from other institutions Vanderbilt Kaiser National University Singapore of all implemented CCP are very similar programs we've been spreading information about this through a Medicare funded learning collaborative and we've just been working recently to create a new not-for-profit organization we're going to call the Comprehensive Care Institute to support our engagement with other institutions implement interested in implementing and studying this because there are literally you know half a dozen or a dozen other institutions we've talked to in terms of policy CMS has been reviewing a physician-focused payment model to try to pay for these ongoing connections to the doctor and we've also been trying to begin to explore what could be done to improve risk adjustment for high-risk patients one of the sort of great tragedies of prospective payment is that it does not do a good job of rewarding you for caring for the sickest patients this is a program that selectively attracts the sickest patients so it's very easy to lose money caring for those patients even when you're in fact saving money on the care of those patients and that is a really sort of fundamentally dysfunctional aspect of the U.S. healthcare system it is somewhat less of a problem and less competitive market context so we're also thinking about trying to spread CCP there although the lack of competition itself has its own problems in those settings so let me just end with going back to this idea of nudge or nudge in some sense it's a false dichotomy you can do both and they may indeed complement each other but I do want to point out to the extent nudges draw resources away from person-centered approaches the trade-offs may be real so when the University of Chicago says it's going to go have a big digital health initiative that doesn't happen without money okay it's a decision not saying it's a wrong one but it's a decision I would argue that for medically complex chronically ill and socially complex patients there are strong reasons to believe that nudge the relationship will dominate nudge and you know it may be for healthy and wealthy people relationships or convenience maybe all that matter and generate resources that may well pay but in terms of making health differences for the most needy I think it's nudge now this is not all just about the doctor-patient relationship relationships with and among members of the team are critical and it's you know they're the key relationship could be with a nurse it could be with a social worker I don't care what I care is that it's a relationship and that it be durable and appropriate for the patient's needs now I will say that I think this idea that relationships matter has profound implications for institutional clinical organization and governance that imply an emphasis on the clinical micro system what does all that mean it means that as a bunch of professionals whether we are doctors or nurses or social workers it's our relationship with each other in the care of patients that really matters and an organizational structure that brings us as teams closer together rather than trying to solve these problems at the level of nurses social workers doctors is probably the one that's going to win in the long run both for all the professionals involved and for the patients and I think all this really implies the importance ultimately of genuine and productive partnership with patients and community that ensure that they're all really part of the team and working together so let me end there and I'll just leave this figure floating over of us reminding you of whatever you happen to feel guilty about so among the many people of the ccp team who are here as many to ask you is right over there who is in fact taking care of young adults and and doing this and helping them manage the transition and we think that that's a really wonderful model let me also just say one of the other models that we're beginning to think about is what we're calling a rounder model where people don't have to necessarily give up their primary care doctor but where the ccp slash hospitalist partners with primary care doctors so that every time you admit you're admitting to the same person right and then you get continuity on that and I think that's a model that has tremendous capacity for both that actually can be combined really well with the ccp model yeah okay I want to initial study with the ccp to the control group what was the difference in number of visits per patient that's the first question the second question have you created relationships with churches but that's also a natural way especially in the black area way of getting residents involved in these projects okay okay god I hope my short-term memory is good enough to get all three okay I remember don't tell Vinnie and no one else either okay so the first one is ambulatory visits which went up I don't remember the amount I'm embarrassed but they did go up not a ton not nearly enough to offset like maybe 10 percent or something like that I do think they concentrated more in the ccp so there might have been a little less fragmentation of it okay so that was the first one the churches so we have worked with churches Reverend Brazer was someone who he interacted with earlier I'll guess I've forgotten his name got a phd from the University of Chicago yeah yeah so so this is this is something that is always on our list we I have two wonderful young research coordinators who work with me who are from the neighborhood and they've gone out and even know Alicia and Felicia and and you know really reached out into the churches and worked with them and so that's been an important strategy for us we are just gearing up for this new study when we did the really big study before we were out in churches a ton and now that we're getting the team trained up to kind of recruit here we'll do that next and the third one was residents and so the answer is absolutely yes I think there is a whole way to rethink how we change residency training to give people more continuity experiences there are obviously challenges which you know better than anyone else in the room about meeting the needs of accreditation and so on but we think there are lots of opportunities we want to as we've already been offering electives for outside medical students to come and visit we are beginning to talk about how to set up resident electives here we're already playing a role in the continuity curriculum for the medical students but I think it's an extremely exciting idea and I think there are a lot of ways to fit residents into this that could be really really good and I'll add not just medicine I think these same issues can work in many many other areas yeah yeah medicare only patients or has that changed so it's it's medicare a and b and having been hospitalized once in the past two years we also some will screen for patients who are actively in the emergency room at any point in time and if we find them there we'll we'll try to recruit them we've taken patients on a sort of humanitarian basis outside of those eligibility criteria if they're not eligible for the study I would love to do it for medicaid patients the challenge is that the institution loses so much money on those people that if we attract more of them to the u of c it creates challenges for the institution's finances we have been talking to hospitals elsewhere on the south side of Chicago that have somewhat better business models for medicaid than we do about ways to do that vinny has also been able to take care of some medicaid patients when in partnership with friend family and so we are doing the most we can do within the institutional constraints that we have and I just do want to say and I do not totally blame our institution I mean I ultimately blame our governor and you know our legislators who have failed to finance these things adequately so medicare advantage is is an area we would love to go into it has exactly the same problems I think some of you are aware of you know some of the growth of medicare advantage in general but particularly here on the south side I will just say that behind a lot of those clinical efforts are in fact large insurers and that those large insurers are willing to subsidize those clinical efforts because the way I believe they are making money is by risk selection medicare advantage like most risk-based contracting in the united states is completely broken okay and until the law changes which it may not change soon or we get you know dramatically better at risk prediction we need really clever strategies to think about recruiting populations as opposed to sort of individuals and there are people here thinking about it but there are not solutions that I think anyone is convinced right now will definitely work but it's critical that we develop those in the ccp model what method or criteria did you use to stratify the complex patients you anticipated with a longer stay versus the simpler patients and then is there flexibility what if someone who went into the category of a simpler they thought this would be like a one night or two day admission and then they actually have a lot of social needs and would require that comprehensive care so again the only criteria in was medicare a and b and hospitalized once in the past year that's it we don't limit it for high levels of social need okay we took all commerce when we did the analyses afterwards we stratified by things like level of measured need okay and we found in general the less activated patients benefited more we didn't find that the patients at higher risk of hospitalization benefited more necessarily um oh there's a little bit of a trend there okay great in your view will crc4p strengthen weekend or change the doctor patient relationship I don't think it weakens it I think I think that it it I you know um Nicole's here she can talk about our our social work program but my sense is that they're compliments that they work together very well and that um um and that the ability of the team as a whole to address needs better makes the patient feel better about the doctor and makes the doctor feel better about the patient and that in the end it is it is not a threat what I do think is a threat is fragmented care from any professional that just adds another person to the mix and then disappears because it reinforces distrust and and um and undermines exactly what we're trying to produce before we uh congratulate that i don't know sir I just want to say that our next session will be two weeks from today on october 16 um john lanterns will speak to us one of the leading pediatric ethicists in the country on on the pediatric doctor patient relationship um but david was a fabulous talk and i want to thank you so much thank you thank you