Mark McClellan: An Economic Lens on Healthcare

Harlan Krumholz: Welcome to Health & Veritas. I’m Harlan Krumholz.‌

Howard Forman: And I’m Howie Forman. We’re physicians and professors at Yale University. We’re trying to get closer to the truth about health and healthcare. Our guest today is Dr. Mark McClellan, but first, we always check in on certain current or hot topics in health and healthcare. What do you got, Harlan?‌

Harlan Krumholz: I sort of tripped on an article this week, Howie, that I thought was kind of interesting and I thought, I don’t know. It’s worth bringing up to you. So I don’t know. Do you like these slushies at these 7-Elevens and convenience stores?‌

Howard Forman: Not for a long, long time.‌

Harlan Krumholz: I could just see you walking out with like a big...‌

Howard Forman: I’m not sure I ever really liked them.‌

Harlan Krumholz: ...48-ounce slush. I’ve seen you before with an ice cream cone that looks like about three feet tall.‌

Howard Forman: That’s true. I love my ice cream.‌

Harlan Krumholz: Yeah, but you’re not a slushie fan, huh?‌

Howard Forman: Not for 40 years, 50 years maybe. I don’t know.‌

Harlan Krumholz: Oh, really? All right. Well, look, there’s an article that came out in the Archives of Disease in Childhood that described a series of cases in the UK and Ireland where children mostly under six years old became acutely ill after drinking these slushie drinks. And so what happened to them? We’re talking about a cluster of symptoms that really kind of mimicked rare metabolic disorders. I mean, it was really odd that they were able to assemble this group and that they were all associated with symptoms appearing after drinking these slushies, and what were they, just decreased consciousness, the kids suddenly became very lethargic, even some unresponsive. Hypoglycemia. They had dangerously low blood sugar levels. Metabolic acidosis. So what’s this? A buildup of acid in the blood, which can be, as you know, quite serious.‌

Howard Forman: Very dangerous. Yeah.‌

Harlan Krumholz: Pseudohypertriglyceridemia, an apparent spike in triglycerides that was actually a misreading. Why? Because these people actually had high glycerol levels.‌

Howard Forman: Right? So tell us about glycerol.‌

Harlan Krumholz: Well, I’ll get to that in one second. And they had low potassium levels, hypokalemia, which can affect the heart and muscles. And the thing was, these kids weren’t sick before. They didn’t have any underlying known metabolic disorders. And in every case, their symptoms start within an hour of drinking one of these slushies. And you know, I mean, everyone listening must know what I’m talking about. It looks like slush and it’s flavored and so forth. So the culprit here at the end of the day, they felt, was glycerol. So what is glycerol? It’s a food additive used in a lot of things. Baked goods, dairy products, even some medications. And in slushies it’s there to maintain the slush texture without needing a ton of sugar. I mean, otherwise it turns out you need a ton of sugar to sort of create that consistency. And what they did was in the effort to try to reduce the amount of sugar in these slushies, they were using glycerol.‌

Now the food agency, the food authority in Europe, has assessed glycerol many times and they’ve concluded it’s generally safe as a food additive. Now, know that the studies on this probably are not so deep, and what’s becoming clear now is that young children may be particularly sensitive to it. The concern is that glycerol in high doses can behave somewhat like a diuretic pulling water out of cells. It’s also rapidly metabolized in the liver, which can be overwhelming in small children, and it can affect gluconeogenesis the production of sugar. And so essentially what they’re saying at the end of the day was, these kids that were just going into a store to drink a fun thing, get one of these slushies, were suffering from glycerol intoxication. And so people are starting to say, should there be guidelines recommending that children under eight avoid slushy drinks containing glycerol? I mean, you might ask me how would I even know that there’s glycerol in it? It’s not like that says slushy with glycerol.‌

Howard Forman: But there is some hint about this. When I was reading when you told me about this, I had no knowledge about this, but I did a little quick reading on it. The reason why it seems like this is popping up in the UK is because there there is a sugar tax. And so if to sell the slushy with full sugar would cost considerably more, and so by inserting the glycerol in lieu of sugar, they can sell it more cheaply. Parents are always sort of fond of something that’s lower in sugar. And so you have the unintended consequences of a sugar tax. I think the reason why we’re not seeing it so much in our country is I’m not even sure too many people are drinking sugar-free slushies. The first question people should ask is, should we be giving the sugar-free slushies to kids?‌

Harlan Krumholz: We need to dig into this a little more. I wasn’t sure they were entirely sugar-free. So it’s not like they said it’s sugar-free, but they’re lower sugar.‌

Howard Forman: Low sugar. Yep.‌

Harlan Krumholz: And I think we need to look at it. I will tell you, having young grandchildren, I think I’ll tell my kids I would avoid the slushies for a while. The thing is, you can treat this. The hypoglycemia can be treated with intravenous high-dextrose fluids. The lactic acidosis, you can use bicarb. There are ways to treat these things, but they’re scary. And of course, are there any long-term consequences? Hasn’t really been studied so much. I would have thought we would have heard about this before if it had been occurring in any sort of frequency previously. So maybe you’re right, Howie, this new thing. But anyway, it’s something to dig into. I thought it was an interesting article, something to stay alert about. Let me just say this, it seems real. It seems real.‌

Howard Forman: It does seem real, but it’s also, it’s a fascinating thing. I’m curious how much we have these low-sugar or sugar-free slushies in the United States, but it is so important that parents are aware of this.‌

Harlan Krumholz: You won’t find me drinking one next week. I’m not drinking that.‌

Howard Forman: No way. No way. Never. Okay.‌

Harlan Krumholz: All right. Hey, let’s get to our interview today. Mark McClellan. He’s going to be terrific.‌

Howard Forman: Yes. Dr. Mark McClellan is the director of the Duke-Margolis Institute for Health Policy, which produces practical policy solutions to improve healthcare value. A physician economist by training with extensive leadership experience, he has previously served as commissioner of the U.S. Food and Drug Administration, administrator of the Center for Medicare and Medicaid Services, and as a member of the President’s Council of Economic Advisers, something that has been done by no other individual in history. He is a member of the National Academy of Medicine and a research associate of the National Bureau of Economic Research. And before his government service, he was a leading scholar of healthcare economics at Stanford on the faculty before it was actually hip to be a health economist and physician. Currently, he’s an independent board member for Alignment Healthcare, Cigna, Johnson and Johnson, PrognomiQ, and is an advisor for Arsenal Capital Group, Blackstone Life Sciences, and Mitre.‌

I’m going to start off with just a background for our listeners. Whenever somebody asks me about my origin story, it always starts with you, because you and I had this phone call in April of 1992 when you were a med student and I was already a radiology resident. I know people look at me and they think that I’m much, much older than you, but we’re basically contemporaries. We had this phone conversation and you’re the one who gave me the advice to pursue the career that I actually have followed to this day. It’s remarkable how much advice you gave me at that time. Now at that time—‌

Harlan Krumholz: What did he actually say to you, Howie? I’m really curious about this.‌

Howard Forman: There are some funny lines in that conversation that I may not want to repeat, but he was basically trying to advise me about whether I should go and pursue a PhD in economics, which is what I was contemplating, having talked to Joe Newhouse at that time, or whether I should proceed down the path that was more akin to where I was, which was halfway through residency at that point. And he told me that getting the MBA and doing the RWJ program at Penn was probably one of the best paths that I could take to achieve the goals that I had because—‌

Harlan Krumholz: Even though he was a PhD, he told you—‌

Howard Forman: No, but he asked me very pointed questions. It was really important. He asked me pointed questions. Did I want to be a scholar? Was I interested in inquiry for inquiry’s sake or was I more interested in being somebody who could avail themselves of policy and coordinate and help people? And he gave me the right advice at that time, and I’m forever grateful. And we met again when he was on the Council in the White House, and that was in 2001 and we have stayed in touch ever since. So I really appreciate you being here. But I would love to hear you tell the story of your career and how you really took the road less traveled because before you, there was Alan Garber and maybe there was someone else that I don’t know about. Alan Garber is now the acting president of Harvard. Before you that was the only MD health economist that I knew about. Now there are a few dozen out there, maybe even more. So tell us a little about what prompted you to go down this path.‌

Mark McClellan: Well, first off, Howie and Harland it’s great to be with you. And as you pointed out in that retrospective, which I think embellished a little bit with the rosy-colored glasses of history, it’s been great working with both of you over the years in policy context and academic context. One thing I cannot take credit for, though, is your podcast success. So I probably got a lot to learn from you all, including on this one about how to do that effectively. My own background was really accidental. There was nothing systematic like the discussion we’re having now. Basically, I loved medicine science and making a difference in people’s lives, and I wanted to go into a career actually coming from a family that was largely made up of people with backgrounds in things like politics and law. I wanted to do something completely different.‌

So I did research in neuroscience as an undergrad, went into kind of an MD/PhD track in that direction, but before too long in graduate school found that that kind of research was probably not my long-term career path. And through basically a series of accidents ended up working in public health epidemiology to sort of pay my tuition. I’d had a bit of background in applied math and just a little bit of background in economics. Found out from that that there was a lot of interest back in the day in understanding rising healthcare costs. And that helped me get a job as a research assistant doing an economic analysis and loved it.‌

Ended up doing a PhD that involved applying data to show, using sort of basic economic statistical methods, that there were some important factors like where you lived and how your doctor got paid that influenced how you got treated, separate and apart from your health characteristics. That was kind of interesting in itself, but applying my medical lens, also used those kinds of factors as, I guess people call these days, “natural experiments.” Maybe not really “natural” and not really “experiments,” but idea that factors that don’t directly relate to your health but influence how you get treated can give insights about how those treatments then actually impact people’s health outcomes and medical costs and things like that. And that’s what got me going. That also led to a job first in the Clinton administration and the Department of Treasury.‌

I had known people like Larry Summers, who was deputy secretary at the time, from graduate school, in the Clinton administration. The Treasury Department was kind of the adult supervision for all of the really heartfelt ideas that HHS and the Department of Labor and other parts of the administration wanted to implement. But the president was a big believer in triangulation and applying economic principles. That led to some experience in Washington that I really enjoyed. And then when President George W. Bush was elected, I got a call and asked to help out just on a temporary transition project around making lower-cost drugs available to low-income Medicare beneficiaries.‌

And like many things in life, Howie, one thing led to another. I really enjoyed my time in government. As you mentioned, we met or spent time when I was on the Council of Economic Advisers, then moved on to FDA and CMS. And since then, tried to find this or try to help advance this space, just like you’re talking about, to make it easier for people who have expertise and interests in certain areas, but who recognize that healthcare and healthcare reform are complicated to find ways to make a difference. I think it’s fair to say all three of us on this podcast right now are trying to do that in one innovative way or another.‌

Harlan Krumholz: You know, Howie, one of the great things about the career that we’ve had is we get to meet extraordinary people like Mark, and he’s so humble, even as he’s talking about this journey it really doesn’t even begin to scratch the surface about what his contributions have been or even just to know that when he was in academics, he produced landmark articles. I mean that JAMA article around distance to cath labs, I mean, is one we still refer to and—‌

Howard Forman: Which fed your very important work on door-to-balloon time. I mean, it all connects.‌

Mark McClellan: Yeah.‌

Harlan Krumholz: He really made a remarkable contribution to then move to service and is widely regarded as someone who’s just had a steadfast commitment to enhancing health systems through evidence-based policy and innovative reforms. And he’s one of the few individuals I know that has such bipartisan support in the era where really serious people were being called to go to government and do good work. He was one of them. And at a very young age was able to manage complex agencies in a way that’s highly unusual. Anyway, I just want to say that you really—‌

Howard Forman: Absolutely true.‌

Mark McClellan: I have one more point here. You said that when I was in academics—I still like to think I’m still an academic, at Duke University. Not Yale, but—‌

Harlan Krumholz: I was more talking about that period before the transition into government service. Of course then you returned to academics, where you continue to excel.‌

Mark McClellan: But I guess the point I wanted to make that I think is going to be relevant for the topics we discussed further is that academics is evolving. And as you all are demonstrating in your programs at Duke as well, thinking about new ways to break down traditional academic silos for the good of solving these really complex and often partisan and heartfelt health policy challenges is very important. This is another shout-out to Yale. I had a great meeting just yesterday with Megan Ranney, who’s the head of the school of public health there.‌

We’ve been collaborating on what the future of public health looks like and can we take advantage of everything that’s going on that both of you are involved with around improving healthcare, data, and so-called “real-world evidence,” not just to make healthcare better but to inform and support healthier communities and really provide a new kind of approach to how public health agencies function in collaboration with a more population-focused healthcare system and with better information coming from electronic health data on diagnostic tests and disease burden that’s showing up in healthcare organizations. Really strengthening what might be a new approach to public health as well. So it’s time for, and there’s a really challenging time in academics, but time for challenge begets reform.‌

Harlan Krumholz: First of all, I realize my introduction to you sound more like a eulogy than a contemporary characterization, but let me just say that... so let’s use your analytic skills. One of the things you were known for more recently was that you were really at the forefront of national efforts to combat COVID-19. You authored response road maps, you co-authored comprehensive papers on strategies for vaccines, testings, and treatment. I mean, you stepped up to this moment. I would like to think we and others were also trying to make our contributions, to be proud of it. I’m an outcomes guy, I’m an outcomes researcher. I care about in the end of the day, what have we really accomplished? And I was looking recently at this Peterson Kaiser Family Foundation health system tracker, and they were showing years of life lost during the COVID pandemic in the U.S. and peer countries, 2020, 2021. And they looked at both the people under 50 and in the elderly.‌

But let’s just look at the under-50 for a minute. Per a hundred thousand people years of life lost during the COVID pandemic in the U.S., we were at about 2000. Now if you look at peer countries, take UK, Germany, and Canada, which were the ones that are closest to us among the peers, they’re about 700. If you look at Sweden, Netherlands, France, Switzerland, Belgium, Austria, they’re about 200, a tenth of where we were. And you look at Japan, by the way, 21. I mean, so let’s just say Asia, okay, we’ll think of them separately. But you think, it’s some of the peer countries—our actual performance was horrific. And yet, what’s interesting is in the current debate people are saying we shouldn’t have even done anything. Lots of people, some people—‌

Mark McClellan: Well, I look at countries like Sweden that did have sort of less lockdowns and so forth. And as you just pointed out, at least for younger people, had better numbers.‌

Harlan Krumholz: Better outcomes. So putting on an academic hat, what have we learned from this and what should we take from the learnings? Politics aside, what are the real hard-nosed academic learnings here that we can apply to the future?‌

Mark McClellan: Yeah. I want to be careful about overstating this because I think one of the learnings has been that in big issues like this where there is a good deal of scientific uncertainty, it’s very important to try to reduce that, but it’s also important to convey that knowledge is going to continue to evolve and the level of certainty along the way involves some uncertainty as well. I think if you look back, you mentioned some of the work that we did, we had a lot of close collaborations with the Trump Version 1 administration and Trump 45 and the Biden administration on things like the ideas behind Operation Warp Speed, which went to the usual way we do innovation in science is you get the good basic discoveries, but then there’s a costly and time-consuming translation process to do the randomized clinical trials to prove safety and effectiveness, especially for products like vaccines that you’re giving to healthy people where even rare side effects could be relatively important, especially those at lower risk, like those under 50.‌

And then once you get the trial results, well, then you start manufacturing and there’s typically, even after trials are done and a drug gets to market, it’s like seven, eight, 10 years before it really starts being widely used. That’s why we’re having a debate around the drug pricing legislation, the Inflation Reduction Act, drug pricing legislation where pharma companies are complaining that, hey, nine years before negotiation kicks in, that’s too short for drugs that aren’t really reaching their full market potential until around year nine or 10. Well, that was clearly not going to work in the COVID context. So there was a concerted effort to try out a different kind of drug development, medical product development, vaccine development, diagnostic development model that involved identifying the key needs.‌

And then doing something more like an advanced sort of volume commitment, figuring out not “Can we wait and see if this treatment works and then toss it out on the market and have a high price and maybe limit uptake and see what happens after seven or eight years as price comes down and the healthcare system adopts to using it?” but “Let’s flip that and have an idea upfront, how many people could potentially benefit from an effective diagnostic or a therapeutic for people who are infected or especially a vaccine?” And come up with a way of doing an advanced sort of volume purchase guarantee that if a product meets these conditions for being safe and effective and appropriate for use, then there’d be a path to purchasing it. And because the queue... remember, the revenue’s for a product developer, price times quantity. Because the queue there was so much higher than would have been the case under the typical product development approach, the price could be a lot lower.‌

And this was special circumstances, a global public health emergency, a pandemic. So people were willing to try more things they might not be willing to try normally. That’s a much different model than what we have now, which is high price/low quantities, as opposed to low price/high quantity. I think where things got a bit off track, though, was in sort of that last mile, we also had a number of recommendations back in the day as the federal government through Medicare and Medicaid and our public funding programs put a lot of relief into health systems, which really needed it to get through COVID a lot. We had encouraged using that as an opportunity to switch in the same way in the care delivery system. You move away from paying for visits and paying for specific services to, hey, let’s track. Especially around COVID, what are the outcomes that really matter in the access to treatments and so forth.‌

And there was some of that, Harlan, but mainly what happened was it was a public health emergency, so money went out the door. We adapted our fee-for-service systems to do things like virtual visits and hospital at home, but there wasn’t really quite as fundamental of a change to how do we get into connections to people. And what really seems to matter, I think, for our population-level health, which are these bottom lines that you were talking about, is not just the medical technologies. I mean, it’s always been true, it’s the choices that people make, it’s their actual access to and use of the medical technologies. And that’s where I think we still have some real work to do.‌

I wish we had had a true deep bipartisan commission that could look at what happened in the U.S. with COVID, just the same way we did—you may remember when I was in government at the time; I think, Howie, you were too—after 9/11 and after the bioterrorism threats and concerns that arose after that. That led to a fundamental look at what went wrong in terms of siloed information and failure to connect risks with public safety protections, a fundamental reorganization of government with strong bipartisan support to try to address those underlying issues. I unfortunately don’t think we’ve done that yet, based on the COVID experience and just still some different camps and different lessons and we’re going to continue to have to work through that.‌

Howard Forman: Can I pivot? It’s not even a pivot because it’s so related to that, but you have been attached in some way to Medicare Advantage from its inception. I mean, you were, I think, on the council when it first launched, but you basically were—‌

Mark McClellan: Well, not quite. I was on the council when the concepts came together. This was a top priority for President Bush. One thing about him was the six things he ran on in 2000, and by golly, he was going to make sure he ticked those off. One of those was the getting a drug benefit and, along with that, modernize private choices into Medicare. That obviously got disrupted fundamentally with the 9/11 attacks, and we had laid some foundations. When I was at the White House and dealing with all of that, then I went on to FDA and I was actually at FDA when the law passed and having a great time there. It’s the world’s most interesting biomedical job. It’s like everything that matters to human health is coming right at you, all the latest science. But the president called and asked if I could go to CMS and manage the implementation of the MMA. So I got kind of thrown in along with a great team there to implement that.‌

Howard Forman: So there’s probably nobody that has followed the scholarship as well as just the reality of Medicare Advantage. And Harlan and I frequently talk about the news about United Health Group or Cigna or Humana with regard to Medicare Advantage. To this day, when I teach about it, I am a typical sort of economist, where on the one hand/on the other hand. Give us an elevator pitch on what is good and what is bad about Medicare Advantage right now and what we could be doing better. Because I know you’ve been critical of it and you do think there are things that we can do better. So I’m just curious from a listening—‌

Mark McClellan: I’ve been supportive of it, too.‌

Howard Forman: That’s right.‌

Mark McClellan: I take personal responsibility for this one because it was our team that helped develop the law and that took the law and turned it into Medicare drug benefit delivered through competing private plans and, along with that, these more comprehensive private plans in Medicare as an option for people to get their Medicare benefits alongside traditional Medicare. And I’d say—I’m going to sound like an economist—there have been positives and negatives. The long-term fundamental trend I think people should keep in mind is that health plans offer one way of addressing a fundamental challenge in healthcare, which is when Medicare was set up and most of our traditional healthcare payment institutions were set up, there wasn’t much that healthcare did for you besides... and on the preventive side, it was all about treating illnesses as they developed and progressed. It was about hospital care. It was about physician office visits.‌

But if you think about care today, and even 20 years ago, projected forward 10 years from now, it shouldn’t look like that. Healthcare should be much more present around you and integrated in your daily life with wearables, with remote monitoring, with digital app supports increasingly facilitated by AI that you need to make sure is safe and effective. But a whole different kind of access that is totally not built around things like physician office visits and certainly not built around, well, let’s wait till people have downstream health complications like their heart disease progresses, heart attack, cancers that progress and so forth. The whole point of biomedical innovation is to shift that curve to a patient journey that is about early intervention, about the things that really matter for their health. And most of those are not medical, at least in the traditional sense. It’s things like diet, nutrition, exercise, weight, software. Medical technology can make a difference, but isn’t the bulk of it.‌

And so this idea of paying a healthcare organization to put together all of those pieces and be held accountable not for, “Okay, well you had an office visit or you had a hospitalization, so we’ll pay for that.” That’s traditional fee-for-service, but instead have a payment at the person level that was linked to what your health risks are and then what your results are. That was intended to really drive towards innovative care that would be, increasingly, digitally enabled, where our technology is going. Again, the future, I think, is even more like that. And so when we set up the Medicare Advantage program, I had two main questions for the staff. One was, how are we going to make sure we pay more for the plans that attract and keep the people who are highest-risk?‌

So at the time the health insurance lobby was pushing for... They knew they were going to get paid on a capitated basis, but you maybe have a few risk adjusters: demographics, whether you’re on Medicaid or not, maybe a few chronic diseases. But our staff, advised by a number of academic experts, and Harlan, you may not have been involved in that sort of thing at the time, but I know you’ve been involved in a lot of sort of real-world data like this. We’re saying, nope, there are a whole lot of conditions that predict higher or lower costs at the time when people are choosing their plan that we really ought to incorporate so that if a plan’s attracting, keeping a much more costly higher-risk patient, they get appropriately rewarded for that. So that can limit the so-called adverse selection problems.‌

So question number one for the staff, “Okay, that sounds good in concept, what’s the best data available?” And at the time, this is around 2005, the answer was health insurance claims. They were electronic, they had diagnosis information, they had information on utilization based on what fee-for-service coverage included, the physician visits, the hospitalizations and so forth. All right, well, let’s use that. So that was what was adopted then on the risk adjustment side. And the other piece was paying for better results. So there we wanted to get to more transparency for patients, but data on quality and so forth was limited. So we identified some measures of quality that could, again, be calculated from claims because that’s what was available, as well as some surveys of patients that could be done essentially by phone. Basically, in those days, these so-called CAHPS satisfaction surveys, things like, “Do you get timely care? Do you have problems in getting appointments?” Things like that.‌

And Howie, if you had told me then that now, two decades later, we’d essentially be using that same system today, I wouldn’t have believed it. This is after $50 billion in so-called high-tech spending to enable healthcare organizations, most healthcare organizations to adopt interoperable electronic records. Lots of information going on with it. What we’ve done, instead of moving to direct more accurate measures of things like risk factors that people have, instead of whether it’s reporting a claim that you had hypertension, well, you know, you’ve got clinical readings of that. You increasingly have clinical readings that people can take it home and validate. We’ve made the Medicare Advantage doctors, the plans who are practicing in these plans, produce so-called encounter data, where they have to go find information from their electronic record but turn it into essentially a claim.‌

And so what we have now is, not surprisingly in the Medicare Advantage plan, because they’re paid on the basis of these diagnoses, not fee-for-service care itself, a whole lot of effort and financial incentives and infrastructure support... into electronic records to find those diagnoses. But then on the quality side, we’re still using essentially the same performance measures. There’s been a little bit of progress to get some clinically based risk factors into the so-called “STARS measures.” But there too, it’s essentially kind of an administrative process on top of usual care to construct the data and enter the data to produce these measures. So even though on the surface, we’ve got a majority of people in Medicare who are now in the Medicare Advantage program I should add, in the traditional Medicare program there’s been some great innovation around moving away from fee-for-service and into so-called accountable care organizations where the healthcare providers take the insurer out as a middleman. Healthcare providers taking on some of the financial population risks themselves for managing patients. But unfortunately, it’s still used based on those inaccurate, burdensome data approaches.‌

So right now, I think the future is definitely continuing in private plans that are delivering care. By the way, the same thing’s true for Medicaid. Over 80% of Medicaid coverage is now delivered through private plans. In traditional Medicare a majority of that program happened under the Biden administration is no longer fee-for-service, it’s accountable care. But we still have this underlying fee-for-service infrastructure that is really leading to overpayments, to much less accountability and transparency around quality of care than we could have. I think the good news is we are now at a point where we can, really need to do something about it. And for all the disruptions happening in the transition of the new Trump Version 2 administration, I think some really serious interest in fixing these core problems.‌

Harlan Krumholz: Well, I wanted to just follow up with one other thing, which is, you and I are both a fan of actually studying new models, these new payment models. So building on what you’re saying, if the fundamental problem is the fee-for-service system, then there’s a lot of opportunity to experiment with new designs. And every design has a potential for benefit or risk. It seemed like the government did a good thing in 2010 when it established the Center for Medicare and Medicaid Innovation, because what they did was establish a group that was tasked with developing and testing these innovative payment service delivery models to enhance quality and particularly to reduce costs. We’d be fans of that; we’re academics. This is actually going to be studying new models.‌

But now if we look at what’s happened, I mean, over the first decade they spent about $8 billion on various models. It did lead to some reduction in costs, but there was a net increase in direct spending that offset that. Of all those new models, there was really very little evidence that it had created benefit, even with the accountable care organizations. And then since then, the CBO, the Congressional Budget Office, is projecting that these activities may have a net increase of $1.3 billion. It’s failing to identify experiments that are leading to... our friend Mark Fendrick at University of Michigan just had one. Canceled with the VBID group. And because it increased costs, now it could be that it’s because you need to look over a 10- or 20-year arc to really appreciate—‌

Howard Forman: Or perhaps collateral benefits.‌

Harlan Krumholz: Or collateral benefits.‌

Mark McClellan: Spillover effects. Yeah.‌

Howard Forman: Right.‌

Harlan Krumholz: But we couldn’t have asked for, and we’re not going to get it now where the government’s going to... I mean, this is, I hope will continue, but I mean, more new money is not going to come into this. Where do we stand with being able to identify which of these models is going to be best and how do we study them? I mean, what is the next step here?‌

Mark McClellan: It’s a great question. And what you’re referring to, the ACA, among other things created the Innovation Center at CMS explicitly charged with spending some money upfront essentially to support and implement these payment reforms, intended to move away from fee-for-service and aligning payments more directly with the care models. We’d like to see stronger primary care, more coordinated specialty care, more focus on, “Okay, how do we get the best outcomes for each person?”‌

Harlan Krumholz: They all sound like a great idea. This was supposed to be the big thing of Obamacare, right? Yeah, this was Obamacare.‌

Mark McClellan: Yeah, yeah, yeah. If we have time, we can come back to another part of the ACA now looks like there’s a lot of efforts in Trump Version 1 to repeal the exchanges. If you look at Republicans today, it’s more about, well, how do we make the exchanges work better, maybe as an alternative in some cases to Medicaid. It’s certainly not about repeal anymore. It’s another very good policy question for the coming years. But back to the Innovation Center. The early days of the Innovation Center, not surprisingly, were kind of, “Let’s try to let a thousand flowers bloom.” Funding going into specific innovative care delivery ideas, maybe sometimes involving innovative products like digital health.‌

I think what was generally found there was, well, if you pay money directly to get this kind of care reform change and get better outcomes and lower costs, but how the heck then do you scale that up? In a healthcare system that both on the fee-for-service side, and I was just describing on the Medicare Advantage side, are not necessarily designed to really support and reward high-value, better-outcome, lower-cost care. So CMS then moved into more of these kind of five-year, or in some cases, longer models where as you were saying, they tried to set up an intervention group and a control population. Control populations are almost impossible these days with so much going on in terms of moving away from fee-for-service.‌

Maybe it’s not exactly what was in the model that you’re testing, but by a year or two or three in, it’s not clear, as Howard was saying, what kind of spillovers have happened in the other groups and what else is changing in those control groups that if those things hadn’t changed, you might really see more of a benefit from the intervention. And some of those did work. I put them into three categories. Dana Safran now at National Quality Forum has done a good review of this. The ones that have done the best were ones that were really focused on getting to total cost of care accountability away from fee-for-service, so some of the more advanced on accountable care organization models. There was a model called AIM, the Advanced Innovation Payment Model for small primary care practices, who needed some help with the upfront infrastructure to make this work. And because they were primary care practices and not associated with hospitals, when they got savings by keeping people out of the hospital, that went right to their bottom line. That one worked.‌

There were some episode models, especially around kind of whole conditions, like more advanced kidney care for people with, say, Stage 3 or 4 end-stage renal disease. They’re essentially managed by specialists like nephrology groups, getting that money into a person-level payment and away from these low payments per dialysis session. And then people getting turfed to hospitals and not having financial support for adopting home dialysis or higher quality-of-life settings, or really getting ready for transplant. Those models worked; mandatory models worked. It’s very hard to save money in a voluntary program. What didn’t work was just some additional payments including for primary care that weren’t strongly tied to total cost of care, getting specialty care more integrated.‌

And I think you’ve just seen and that one of the first things the new administration did was pull back some of the models in the Innovation Center, Harlan, that, based on the criteria I just described, were most likely not to show savings. Even though they weren’t done yet, they were not on track for it. That VBID model, the model that Mark Fendrick supported, he rightly pointed out that this wasn’t really a controlled experiment, that there are other ways that Medicare Advantage plans are providing benefits that are targeted to individuals based on their risk to try to help them get the treatments or the providers that are best able to manage their conditions chronically. But guess what happened in the model? Some of that did occur, but what also occurred was is, these patients got more engaged. Guess what? They had many more of these diagnoses that are undertreated, coded, and that led to additional Medicare Advantage risk adjustment payments for the reasons that I already described.‌

So what we’ve gotten good at in both the Medicare Advantage side, and don’t leave Medicare Advantage out of this, but also in the innovation transition away from fee-for-service, this direct contracting side, is identifying ways to improve care and lower costs within care. But we haven’t gotten as good at doing that in a way that still saves money overall. The additional risk adjustment payments that came out in the VBID model, the additional upfront payments that many organizations got because it was a voluntary program, the ones who signed up figured out they do better against the risk benchmarks in these models. That’s what needs to be solved now. And I think we’ve learned a lot about how to do it. And I think the big challenge, and this is a timely challenge, we’ve got much better data than we’ve ever had. We’ve got a commitment and more pressure to come up with better ways to deliver care, to lower costs using approaches that are just not paid for in fee-for-service. AI, home-based care, non-medical interventions targeted effectively. Can we do it, based on what we’ve learned so far?‌

Howard Forman: We are really fortunate to have you as a friend, but really for the nation to have you as a scholar and somebody who is able to inform policy decisions without partisanship. And you are doing an enormous service for all of us, even when you’re not in public service. So thank you so much for what you do and what you continue to do. And hopefully we’ll have you come back to Yale maybe in the next year because I think we’d benefit from more time with you.‌

Mark McClellan: That’d be great. This is the most fun I’ve had on a podcast in a long time.‌

Howard Forman: That’s awesome. Thank you.‌

Harlan Krumholz: It’s great talking to you. Hey, that was a great interview. He’s amazing.‌

Howard Forman: We both love Mark. I mean, this is somebody that you want to be in charge of things.‌

Harlan Krumholz: He’s got such a depth of knowledge on things. But let’s get to the Howie Forman part of the podcast.‌

Howard Forman: Very short today, very short. This week is Match Week, when our graduating medical students will find out where they do their residency. But at the same time, we’re also in the midst of medical school admissions season at Yale. And so many other schools are sending out offers and inviting students to campus to take a second, or in the time of Zoom, it’s actually their first look on campus. So hopefully, we’re going to come back and we’ll visit this topic again. But I wanted to comment on a recent paper that came to my attention via our friend and podcast guest, Dr. Art Caplan at NYU. One big question that keeps coming up is, how much medical school does one need? The majority of Yale medical students take five years or more. Most schools are four-year medical schools. And then there are novel approaches for three. And NYU is among those schools. So in this paper, investigators looked at the first seven graduating classes of experience with their novel three-year accelerated medical school curriculum at NYU.‌

And the results were very favorable by essentially all metrics. The three-year students performed as well or better than their four-year compeers. And this was judging by residency outcomes to the extent possible. And as someone who went through an accelerated undergraduate medical curriculum, I did a seven-year bachelor’s in medical school. I’m familiar with the many trade-offs. And by no means is this paper the final say. The students in the three-year track are self-selecting, they’re not being randomized, but it should remind us to keep thinking of ways to reduce the financial burden on our youngest physicians and help improve the quality and size of the workforce and not become sclerotic in our thinking of the right and wrong ways to train new doctors. I’m curious to hear your thoughts on this.‌

Harlan Krumholz: Well, look, just knowing you went through an accelerated program and how great you are, I’m sold. I don’t even need any other studies.‌

Howard Forman: I’m stunted, Harlan. I’m stunted.‌

Harlan Krumholz: I think we all know that first of all, there have been a bunch of accelerated programs throughout the country for a while, but the acceleration was usually at the undergrad level, not so much at the medical school level, but if we all reflect on our medical school experience, I think we all think we could probably have squeezed out 12 months without in any way interfering with our ability to emerge as a physician. And you know, most programs have gotten now fifth year, so people are...‌

Howard Forman: I know. Yeah.‌

Harlan Krumholz: So thank you so much for reporting on this. It bears looking at and with the extended training periods for physicians, we need to find ways to become more efficient.‌

Howard Forman: Agreed.‌

Harlan Krumholz: You’ve been listening to Health & Veritas with Harlan Krumholz and Howie Forman.‌

Howard Forman: So how did we do? To give us your feedback or to keep the conversation going? Email us at Health.veritas@yale.edu or follow us on any social media, particularly LinkedIn or Bluesky.‌

Harlan Krumholz: And we’re always welcome to your feedback, questions. Tell us about your experiences. We always want to hear that. We heard from Josh Gabala this week. Thanks, Josh, for giving us some feedback. Maybe we’ll share that on a future podcast, but everybody, look, we’re very interested in what you’ve got to say.‌

Howard Forman: And if you have questions about the MBA for Executives program at the Yale School of Management, reach out via email for more information or check out our website at som.yale.edu/emba.‌

Harlan Krumholz: Health & Veritas is produced by the Yale School of Management, the Yale School of Public Health. Thanks to our fantastic researchers, Inès Gilles, Sophia Stumpf, Tobias Liu. Then to our remarkable producer Miranda Shafer. Talk to you soon, Howie.‌

Howard Forman: They are amazing. Thank you very much, Harlan. Talk to you soon.‌