From 3M Health Information Systems
Podcast Episode Transcript: Seeing value from the patient’s perspective
Gordon Moore: Welcome to the 3M Inside Angle podcast. This is your host, Gordon Moore. And with me today is Dr. Kenton Johnston. He is the associate professor in the Department of Health Management and Policy at the St. Louis University College for Public Health and Social Justice. He also has an appointment in the Department of Health and Clinical Outcomes Research in the School of Medicine. His PhD training is in health services research and health policy from Emory University. Prior to academia, he worked in the medical informatics field at Blue Cross and Blue Shield. Welcome, Dr. Johnston.
Kenton Johnston: Thank you. Thank you for having me.
Gordon: The reason I think it would be interesting for our audience to hear from you today was because of the viewpoint article that I saw in JAMA, where you wrote about building a better clinician value-based payment program in Medicare with several other authors, let me note. What I’d love to hear from you is, why did you write this article? Where did the impetus come from?
Kenton: Sure. Well, I’ve been doing a lot of work on Medicare value-based payment with actually, the two co-authors in this article, Dr. Jason Hockenberry, who is now the department chair of health policy at Yale University, and Karen Join Joynt Maddox, who is a cardiologist at Washington University and also the co-director of their Center for Health, Economics, and Policy. And we’ve been doing a lot of research on how the Medicare value-based payment program, especially for clinicians, so physicians and nurse practitioners impacts not only their performance, but also if it has any unequal impact, inequitable impact for those who treat more vulnerable populations such as those who treat poor populations or racial and ethnic minority populations or rural populations.
And so we’ve noted a lot of issues with, especially the MIPS, the merit-based incentive payment system program. And there’ve been a lot of people that are just like, let’s scrap value-based payment altogether because it’s just not working out well. And so we were thinking this through, and that’s what generated this article is we’re thinking, well, value-based payment has not accomplished what it set out to accomplish. However, we probably don’t want to go back to the old program that everyone was complaining about, which was just the straight up fee for service model that we saw a lot in the 1990s and 2000s, when we saw a lot of cost growth in payments to physicians and that motivated these value based payment programs in the first place.
There’s a whole background we could get into if we wanted to relate it to the sustainable growth rate payment formula and how that wasn’t actually sustainable, and congress eventually, this is one of the few bipartisan things that congress passed, the Medicare Access and CHIP Reauthorization Act in 2015, which packaged a lot of these value-based payment programs together for physicians. A lot of people agreed, in Medicare and in the private sector, that we want to do more than just pay fee-for-service, we want to also pay for value. And so that motivated us to write this piece and think about, acknowledge what the problems and failures have been, but then argue for what might a better value-based payment program look like. And also, what are some successes that we can learn from as well in value-based payment.
And so that’s what motivated the article. We’ve done a lot of research pointing out problems in the value-based payment programs, but we don’t think that that means we should just scrap value-based payment altogether.
Gordon: Yeah. Throwing the baby out with the bath in the sense. I mean, we go back to the classic fee-for-service, that’s beneficial in a sense because it drives clinicians to do things for people, but sometimes if there’s no clear value in that to the person, there’s some estimates of significant proportion of that stuff being maybe unnecessary, and that may be that could lead to harm unintentionally. And so this drive for value is looking for ways to say, can we line up payment as a signal to let’s hue more close to do the right thing?
Kenton: Yeah there are different ways of defining value, we get into that. But ultimately, I think the three of us would argue that value should be defined by the customer, by the patient in this case. And we want a system that best meets patients’ needs in an as efficient way as possible, but also without skimping on care, we want to achieve the best possible quality and outcomes for patients. And that’s hopefully what a value-based payment system would incentivize.
Gordon: You mentioned that actually in the viewpoint, you and your co-authors, and they talk about, there’s a lack of patient reported outcomes in many of these program metrics. Tell me more about that.
Kenton: Well, a lot of program metrics are based on data reported through medical claims or other electronic systems, and they tend to measure process of care measures, which is fine. If you have diabetes, you should be getting your HbA1c checked, your LDL cholesterol checked, that sort of thing. And that’s all fine, however, if you ask patients what they value most and what they care about, it’s usually outcomes related to functional status, quality of life, staying out of the hospital and staying home and being healthy and being able to live their lives. And it seems like there’s a need for more measures that look at those patient reported outcomes as well in addition to the process measures. And it requires a change of framework, which is seeing things from value from the patient’s perspective.
I know Michael Porter at Harvard University in the business school has written a lot about this in terms of how values should be defined from the patient perspective because they’re ultimately the customer.
Gordon: Yeah. I mean, it occurs to me that we often measure what we can get our hands on, and that’s the classic problem of the drunk guy in looking for his keys under the street lamp because that’s where the light is as opposed to that’s where the keys are. And it seems like that’s one of the fundamental problems we have in understanding value and quality in health care, we have oceans of administrative claims data, and they can tell us a lot.
There’s utility in that, but it can’t tell us everything we need to know. And it distracts us from thinking about how do we answer the question, is this really working for people? Is the goal that we’ve had a visit twice a year for diabetes care, or is it the goal that a person feels like they’re healthy enough to be at work and productive? But we don’t seem to have those data streams.
And the tricky thing about outcomes is that a lot of things impact outcomes. In addition to the medical care provided by physicians, there are things outside of physicians’ control. But outcomes are impacted by biomedical care as well, and so it’s important to measure them in addition to processes of care. There have been studies conducted that show that following processes of care for conditions like diabetes, for example, or heart failure do explain better outcomes in patients.
When I think about these other things, let’s say functional status, a person’s able to do more for themselves and get around better, that seems very logical, it seems like exactly what a person wants when they have a hip replacement, for instance, and yet there’s a burden of work that comes with that. And one of the other issues that you talked about with some of these value-based programs is that there’s a significant administrative burden. So tell me about that.
Kenton: The MIPS program is a great example of this. There are over 400 measures clinicians self-select a subset of those measures that they want to report on. And so in order to do well on the MIPS program, you have to be good at navigating the administrative system and reporting. And this creates a burden on clinicians. There have been some papers actually looking at what is the cost to physician practices of having to report on all these measures, and it’s in the neighborhood of tens of thousands of dollars per physician.
And that’s fine if it results in better patient care, but if all it does is result in jumping through more hoops without really improving patient care, then that’s a problem because that competes with time and resources that could be spent focusing on the patient. And there’s a reason for the administrative complexity because you have physicians in all kinds of different specialties, and they’re going to argue that this measure is more important to them if they’re a cardiologist versus a psychiatrist versus an endocrinologist. So I think MIPS has erred on the side of trying to keep everyone happy, and that’s why there are so many performance measures.
But that also creates a lot of complexity, and that complexity takes time and resources away from patient care, and it also advantages those physicians who belong to large health systems with the administrative and technological capacity to continuously track these measures and game the system, so to speak. Large health systems do this, they identify because their physicians report is groups, track all the measures, identify the ones that they perform best on and report those to CMS. I mean, that’s only rational from their perspective. But that creates an advantage if you belong to a health system versus if you’re an independent practice, for example, in a rural area.
Gordon: I always wondered about that because I’ve spoken with a lot of small, independent practices who complained bitterly about their burden of attesting to various different processes within their practice and trying to extract, painfully, information from electronic medical records to report on quality. And then in larger health systems, it seems like somebody could be designated as the data extraction person, another person can fill out forms and submit, and the cost for those are absorbed within a much larger organization. So there seem to be a pretty significant structural bias in favor of large entities, which makes me think that when I see studies that say large entities appear to do better on quality indicators, it makes me think, wait, wait, there’s a structural bias there. So you you’re telling me that you’ve seen the same thing?
Kenton: Yeah. We had another study that came out in September in JAMA finding exactly this, that if you belong to a large health system, you perform much better under MIPS. And a good portion of that appeared to be driven by measures that were dependent on technology and reporting. And I don’t think it’s any coincidence that as we’ve seen a drive to value-based payment in Medicare and the private sector that you see a lot of consolidation in physician practices being bought up by larger hospitals or health systems.
That really has increased over the past decade because most physicians do not want to spend most of their time on administration. And so from their perspective, it makes sense to join a house system because someone else can take care of all that while they can focus on clinical care. Of course, some of them may not want to do that because you may give up some independence in how you prefer to treat patients when you join a larger organization. But there’s definitely a strong trend towards consolidation of physician practices into larger organizations.
Gordon: And then one of the other pieces you talked about in terms of a problem with the current model is the complexity of risk adjustment. You were touching on that a little bit earlier. You want to unpack that a bit for me?
Kenton: Oh, yeah. This is a huge problem. When you’re trying to compare any health care provider, if we’re comparing physicians, physician A has a different set of patients than physician B, and you don’t want to penalize physician A if he or she happens to be a geriatrician and sees older and sicker patients. Of course, they’re going to have higher costs and worse outcomes. So CMS Medicare uses the hierarchical condition category risk adjustment model that adjusts for comorbidities. And it has all the major co-morbidities in there, but there are certain things that doesn’t risk adjust for like dementia and depression are good clinical examples, it also doesn’t adjust for poverty or education, or the neighborhood that live in.
And those things all impact health care costs and health outcomes and access to care. And research shows that clinicians who treat more poor patients or patients with dementia end up performing worse on these measures because the risk adjustment system, CMS, uses, doesn’t capture that particular area of patient complexity. And so the impact of that could be to penalize safety net practices who treat a lot of low-income patients. It could lead clinicians to avoid complex patients or poor patients because it impacts their performance ratings. And I don’t think those are consequences that CMS would want to have, they’re unintended consequences and they’re caused by an inadequate risk adjustment system.
This affects not just Medicare fee for servants performance measures, but also the entire Medicare Advantage program, because the same risk adjustment system is used there. And so if depression is not included, for example, in the risk adjustment system, then why would a Medicare Advantage plan spend a lot of time and resources on building out provider networks to treat depression and to screen for depression and to manage depression? Prior research has showed that some plans actually engage in service-level risk selection, so intentionally having narrow networks and less services for certain conditions that are deemed unprofitable. Those are again, unintended consequences of risk adjustment. So it’s a nerdy topic, but it’s really important because it impacts it impacts payment, which impacts incentives for treating high-risk patients.
Gordon: Again, it just makes me think we’re all human, so we all have these feelings and we need to have checks and balances in place. And it goes back to Willie Sutton, follow the money. If we’re going to pay for stuff in a certain way, that’s something that the recipients of those payments pay close attention to. How does the money flow? How do I win? Because of course, we all want to win with that. And if the symptoms are weird, we’re going to get weird results. Well, let me pivot to some of the successes in value-based programs that you’ve seen. Tell me, did you guys in the article come up with some that you thought were laudable?
Kenton: I mean, all three of us have been paying pretty close attention to the literature on value-based payment. And there does seem to be consensus emerging from the literature that the Medicare Shared Savings program has resulted in modest cost savings. So we’re talking about a couple of percent here, a few percentage points in savings, so it’s modest. Well, not at the same time, either maintaining quality of care or improving, so it’s not like they’ve done that by reducing quality of care. And probably the researcher is most well-known for the research in that area are the two Mikes at Harvard, Mike Chernew and Michael McWilliams, they’ve both done a lot of research in this area, so we are citing their studies.
And in particular, those savings were concentrated among ACOs operated by clinician groups, so by physicians. And so the argument might be that physicians are closer to patients, and so they may be able to better design accountable care organizations to meet patients’ needs than a hospital system might do. The incentives are a little different for hospitals versus physicians as well. Another important piece of that is, Medicare realized that some of these ACOs were locating in wealthier areas, and so they introduced an investment model in 2016 in order to expand ACOs into poor and underserved areas and urban areas and rural areas, and those models were also successful.
And so that shows that ACOs can work among not just wealthy populations out in the suburbs, but also among poor, rural populations or poor, urban populations who may have less access to care and comorbidities and that sort of thing. So there’s more research that needs to be conducted on a shared savings program where that appears to be an area of success. And then Medicare Advantage, well it’s separate from Fee-for-Service Medicare. It covers the latest statistics I saw were 42 percent of Medicare beneficiaries, we could easily hit 50 percent in the next five years. And that program is basically providing managed care for beneficiaries in a way that the fee-for-service program does not.
So you have all kinds of different payers assuming population-based risk, and some of them are able to be innovative and provide benefits that beneficiaries can’t get in fee-for-service Medicare, including things like extra services and supports to help meet peoples’ social needs. Some of these, and typically people think of Medicare Advantage insurers just as United and Blue Cross, and Humana, which are the largest insurers, but they also include health systems that offer insurance like Geisinger and Henry Ford, or Kaiser is another example, that have experience not just insuring, but delivering care to beneficiaries.
There have been a number of studies conducted on Medicare Advantage, and the studies tend to show the beneficiaries receive similar, better quality of care in Medicare Advantage and have lower hospitalization and mortality rates than traditional Medicare. We have to caveat that research somewhat because there are also findings that show that Medicare Advantage tends to treat healthier beneficiaries than traditional Medicare, but there is definitely been success in that program. And you could think of a Medicare Advantage plan as like a large ACO in that they’re assuming population-based risk for managing the total care of beneficiaries. So that’s another area if you went on a defined value-based payment broadly, success.
And then lastly in the private sector, the alternative quality contract of Blue Cross Blue Shield of Massachusetts, that’s been operating for over 10 years now, there’s been a lot of studies conducted on it. It seems like every two years, there was another one in new England journal of medicine. And this involved, again, population-based assumption of a risk with global spending targets for clinician groups, and then they share in savings and losses. There’s a lot of similarity between that and the Medicare Shared Savings Program, similar type thing. The latest study in New England Journal of Medicine showed 11.7 percent savings in annual medical costs associated with that program.
The other thing I think that’s notable is they’re evaluated on measures of care quality, but the number of the measures is 64. Contrast that with 400 in MIPS, 64 seems a little more reasonable, less of an administrative burden. And also, clinicians were given continuous feedback on a timely basis from Blue Cross Blue Shield of Massachusetts, which is what clinicians want from a performance system so that they can alter, change their performance. If something is happening that they’re not aware of that isn’t good, being given timely feedback helps them adapt and better meet the needs of patients in real time. And you don’t see that in MIPS program and in Medicare, for example.
Gordon: More like you have to wait until the program evaluation period a year and a half later to find out how you performed, that kind of thing?
Kenton: Yeah. So, I mean, the information is going to have less impact on your practice when that’s the case.
Gordon: Absolutely. So from all these—when you look at these different models and they’re subtly different, but there are some commonalities beneath them, what would you recommend? What do you and your co-authors think would be the way to move the country forward with value-based care?
Kenton: Well, they all seem to be a movement away from the fee-for-service payment structure. So in certain versions of the Medicare shared savings program, they’re assuming a large portion of savings or losses and splitting those with Medicare. Medicare Advantage assumes full risk, and the alternative quality contract, they were sharing in savings and losses as well. So one commonality is a movement towards more global payment structures. Population-based payment structures used to be called capitation, people don’t like to use that word anymore, but basically that’s what it is. And so that transfers the focus from particular services to care of the whole patient across providers within a health care system.
And the incentives are different under that kind of model. There would be more incentives to coordinate care across providers, ideally. There would be more incentives to meet the patient’s need. So a good example would be, right now, under the hospital re readmissions reductions program, the focus might just be on preventing a readmission, but then the question is, well, what happens to the patient who is not readmitted? Is this actually better or worse for their care? What happens to their post-acute care under a global payment structure? The system would be responsible for not just the hospitalization, but their post-acute care. And so they’d want to consider the whole patient, what’s actually best for the whole patient.
That’s a benefit of that system. The trade-off would be it would incentivize further consolidation within the health care industry, or at least independent practices would have to form contracts and partnerships with larger systems under that system. But global payment is a common feature of these successful models, real time, quality improvement. So being given more—clinicians getting more real-time data on quality of care and outcomes for their patients, that was a feature especially of the alternative quality contract. A lot of ACOs, the good ACOs are tracking their clinicians and measuring patient gaps in care so that clinicians can fill those gaps, for example, and a lot of Medicare Advantage plans do that as well.
Another feature we called out was organizations led by clinicians and physicians themselves. That’s important because clinicians have more regular contact with patients, and a hospital run organization may not actually have as many incentives to keep patients out of the hospital because from a financial perspective, that’s a loss of revenue for them. And the number of quality measures are fewer when you look at especially the Medicare Shared Savings Program and Alternative Quality Contract, we’re not talking about 400 measures, we’re talking about an order of magnitude less than that, so I think that.
And also, they have mandatory measures. The MIPS Program, the measures are not mandatory. You, as a clinician or practice, you pick which measures you want to report, and that creates incentives to just report the measures you happen to perform best on from year to year and you can’t really compare one, the performance of one physician to another or one physician group to another when everyone’s picking their own measures. So those are some common framework. The other thing we added in there is inadequate risk adjustment is always going to be a problem. We can never develop a perfect risk adjustment system.
And so it may be important to develop incentives for improvement overtime and judging participants against their own historical performance rather than just external benchmarks. And especially, that may be important for practices in poor and underserved areas, and providing them also with infrastructure they need to succeed. So the idea is to help make them more successful and reward them for that.
Gordon: That sounds like a nice setup. I mean, it sounds like if you can provide the infrastructure for practices, let’s say their independence, but they can maybe come together around some entity that has the contracting capacity, data, analytic capacity, and resources, then it might help them do better. If you measure against performance, you’re mitigating at least against some of the risk adjustment problems with social disparities and other things that impact their patient’s outcomes. And fewer metrics, but not cherry picking sounds like a smart way as well.
Kenton: Yeah. There needs to be some mandatory measures.
Gordon: Well, that sounds good. I mean, I’m encouraged in the sense that it seems quite rational, it seems quite doable. Now, as we move to close, I’m curious, what do you think is the likelihood that we move, or how quickly do you think we should move?
Kenton: Well, within the Quality Payment Program in Medicare and that the MIPS Program is part of, and the Shared Savings Program is part of, they call a certain segment of the program, the Advanced Alternative Payment Model Program. And it gets that name because in order to participate in it, you need to share in both savings and losses, upside and downside risk with CMS, which is a movement towards global payment. And in the first year, only 4.3 percent of clinicians participated in that, but in the second year, that percentage increased to 17.3 percent. And part of the reason is they get a five percent bonus if they participate in that program, so CMS is really financially incentivizing that.
So if that continues to grow, then we will see a movement towards more global payment structures in the Medicare Fee-for-Service Program. And again, as I mentioned before, we’re seeing increased growth in the Medicare Advantage Program, which is assumption of population-based risk. And I anticipate that will continue to grow in the future as well.
Gordon: What’s your next research agenda?
Kenton: My next research agenda. Well, I’m working on some projects looking at how well the Medicare program meets the needs of Americans with disabilities. A lot of people think of Medicare as a program just for the agent, but about 15 percent of beneficiaries receive their benefits due to disability entitlement. And they have a unique set of needs, so I’ve been doing research, comparing the fee-for-service and Medicare Advantage programs on how well they meet the needs of beneficiaries with disabilities. So that’s going to be an area of focus for me over the next year.
Gordon: Dr. Kenton Johnston. Thank you so much for your time and thoughts today.
Kenton: Thanks for having me. It was great talking with you.