The Perverse Incentive Trap Hidden Inside Value-Based Care - and What to Do About It. EP512

This file was generated by Descript Episode 511. The Tension When Clinical Teams Take on Risk for Policymakers and Others Looking to Rustle Up Future Perverse Incentives. Today I speak with Dr. Siva and Dr.

Monica Lypson. American Healthcare Entrepreneurs and Executives You Want To Know, Talking. Relentlessly Seeking Value. Last week we talked Medicare Advantage with Betsy Seals, and we talked about finding members who a plan can serve well.

This makes sense because Medicare Advantage is a capitated program. In other words, Medicare Advantage plans get paid by CMS a per member per month, and they have to keep their expenses lower than the per member per month payment that they are receiving. Now they will get more money if the patient is coded at a higher clinical complexity. In other words, for sicker patients, you get paid more.

Therefore, upcoding has become a thing and you see a number of plans, Medicare Advantage plans getting in a bunch of trouble for upcoding. That's a great way to make money if you are a Medicare Advantage plan, because right, you're getting paid to take care of a sick patient who's not actually as sick as you say they are. Therefore, they're not gonna rack up the medical expenses that such a sick patient actually would 'cause they're not actually that sick. Meanwhile, on the other side of the house, we have some health systems upcoding visit complexity just as fiercely.

And then we see the same Medicare Advantage plans who have been in the past, at least busy with their own upcoding, who are now doing automatic down coding of the health systems upcoding. Takes one to know one, I guess. What a time to be alive. Meanwhile, in episode 505, I spoke with Dr.

Ahilan Sivaganesan, Dr. Siva, and we talked about calculating the Operative Value Index. What is the operative value of any given surgery or some other episode of care? And I want to play a clip that I cut out of that episode at the time, because yeah, it was just a whole other conversation and there was a whole bunch of other vectors and conclusions that could be drawn.

So I wanna follow some of those vectors now and foreshadowing, it definitely has to do with this upcoding business. So here's me talking to Dr. Siva. When it comes to surgeons, here's why.

I think time-driven, activity-based costing is existential and like absolutely necessary in the coming years. But you know, when I present this kind of information, this data infrastructure that allows you to do prospective, time driven activity based costing, when I present it to other surgeons and when I present comparisons, Hey, surgeon X is 30% cheaper than surgeon Y, you can predict what the next comment is, right? Well, that's not really a fair comparison. My patients are different than his patients.

What that comes down to is that there are certain confounders that we know, and it's very procedure specific, specialty specific, that I know as a spine surgeon, what are the key patient specific and surgery specific variables that account for these dramatic differences in cost. And so you have to incorporate those confounders into the analysis so that when you do a ranking or a comparison from one practice to the next, or one surgeon to the next, you are accounting for all those potential variables that could be at play.

And that's what an employer would also want, right? If they, if they wanna make an intelligent decision, we want to go at risk for spine care. We're trying to choose between these two, quote unquote, high value practices that are ready to play ball with us. And we wanna understand in a very quantified way what the costs are, what the payments are gonna be, and what the outcomes are.

You have to have access to those confounders and be able to adjust for them. So this is a critical foundational requirement for any type of value-based marketplace. And so I think it serves multiple purposes. So like relative to cost accounting and there is a methodology called TDABC or time-driven activity-based costing.

Again, TDABC. But what we talked about here is just A, the fact that this doesn't exist. And then B, there's two ways it could go if it does exist. One is that you could use it for games, right?

Like you could use it, which by the way, is one of the main reasons not to bring up a sore subject, why it became illegal for doctors to own hospitals. Because doctors were getting real, real good at cherry picking and lemon dropping, and then got kicked outta the C-suite. On the other hand, though, not knowing this information, recognizing that. You're losing money in ways that you don't have to be right, like you do not have the action-based insights to actually take action if you don't have these kinds of things.

So this is bigger than surgical outcomes, but certainly entwined. And I think the point that you're making is, other industries have done this. It's complicated. Sure.

But we've seen the downsides of not knowing the information. We need to figure out how to move forward in the spirit of the endeavor. I totally agree. You can never expect a physician to go at risk, in any way, if they don't understand their own costs.

This is a prerequisite, right? And so if you, if any, anyone who believes either in the short term or the long term, that physicians going at risk for outcomes and costs is necessary for true movement towards value-based care, must understand that physicians have to know their own costs. How else can you go at risk? Otherwise you're jumping blind into an abyss.

And so that's, that's the way I think about it, is that knowing your own costs in a prospective scalable condition, specific way, is a absolute minimum requirement for physicians to go at risk and, and move us towards a value-based marketplace. We know that procedural bundles are coming, whether it's Medicare and the latest CMS program or the commercial payers following suit. The payer is gonna come and say, alright, you want to do this elective spinal fusion? Here's how much money you get and this needs to cover your 90 day global, or whatever, whatever that may be.

If I don't understand the true cost of delivering care for my practice or hospital, within that time horizon, I'm gonna get screwed. I won't even know which type of patients to pick. To include in that bundle because there's gonna be a sliding scale, right? There's gonna be the super complex sick patient that needs a revision third time operation that's gonna be a bundle buster where I'm gonna lose money on that patient.

And then there's gonna be the other young, healthy, first time operation patient where my costs are gonna come under that predetermined lump sum from the bundle. If I don't know how to predict which patient is favorable for a bundle or not, I'm gonna be in no position to be able to enter the fray of any type of at risk contracting, whether it's at the procedural level or at the condition level. Okay. So, you just operationalized cherry picking and lemon dropping.

Which I'm sure wasn't your intent. Right. What happens with those patients who need that third revision and Yeah, it's a great question. One thing I've been talking about for a while in our published works is the idea of sliding scale bundled payments.

So if you have the ability to do time-driven activity based costing, and you can show a bell curve of true costs based on the complexity of the patient and the complexity of the surgery. Then you're armed with the information to go to a payer or a self-funded employer and say, here's the distribution of my true costs. So here's what I'm looking for from a distribution of actual bundled payments that reflect the complexity of the procedure and the patient. So you know what?

We don't wanna avoid taking care of that super sick patient with the needs of fourth time operation, but based on our data that we're gonna make transparent to you, our costs are about 35% higher to take care of that patient. So we're asking for a payment that reflects that difference, so that we can take care of the whole pie. But the only way you're gonna be able to do that nuance and have that sliding scale bundle payment is if you're doing time-driven, activity based costing in a regular basis prospectively, and then be able to slice and dice it like this.

I brought up the cherry picking and lemon dropping. In short, it's when. If there's risk-based anything, there's a lot of incentive to pick up patients, i.e.

cherry pick patients who are gonna do real well and not cost a lot. And then lemon drop, in other words, if somebody you think is gonna be really difficult and be costly, you're kind of like, oh, we don't have time for you today. You get rid of them. And what you're saying though is you sort of as a practice have to know what your costs are in order to effectively run your practice and then to help make sure that lemons don't get dropped or cherries don't get inappropriate operations.

You can use it to negotiate to stop the need for that to happen to begin with, and you're doing it in a very informed way. Now, this will require someone leading an organization who is quite interested in the spirit of the operation and not revenue maximization. Yeah, no, there, there's, you gotta avoid the gamesmanship to the point of conflict of, of interest. I'm glad you brought that up.

I sometimes chuckle when I hear this notion that, Well, you know, physicians shouldn't own hospitals because then they're just gonna steer patients in a certain way and game the system. But you know what my fundamental most deep conflict of interest is right now is that I get paid to do surgery. It's fee-for-service. That's the deeper conflict of interest.

Interesting. Right. Dr. Siva summed up the tension here really well.

On one hand we say physicians should own hospitals and shouldn't cherry pick or lemon drop. On the other hand, we say it's all about providers going at risk and built into that going at risk is selecting patients. So the very, you know, holy grail that we are promoting here, i.e.

providers at risk in value-based care has patient selection and risk incentives baked into it. While we simultaneously claim that we can't trust doctors with these same incentives. Conundrum, big one that we just uncovered here, and it's crazy to me that we are this far along into the value-based care thing, and I do not recall anybody saying this directly in the way that is coming out in this podcast episode right now. If we think that we're gonna pay health systems in particular, we're talking about big corporatized companies, organizations here.

If we think that we're gonna pay these organizations on a sliding scale, how is that really any different than basically handing them their own version of a RAF of a risk adjustment framework? And we know exactly what happened? We just talked about it. What happened with the Medicare Advantage RAFs.

Again, we're not talking about individual doctors. These are big organizations with C-Suites that are incentivized on growth with whole departments dedicated to revenue cycle management or whatever we're calling, maximizing the revenue and the profit off of any given transaction for good reasons and not so good ones, just to be clear. But what is definitely pretty unarguable, everybody has a very vested interest in early and often the highest codes possible. That is what the incentive is, so it would follow that if we hand out a RAF, risk adjustment framework directly to these health systems in the form of a sliding scale, look, in T minus 10 minutes, the same health systems with these separate departments held bent to upcode, like every ER visit to a level four complexity, listen to the show with Al Lewis are probably going to, odds are good, they're going to upcode almost every patient some material complex level.

Kalshi or Polymarket probably won't even let you bet on that prediction because it's so obvious that this is what's gonna happen. So many patients, so many members will achieve lemon status and health systems will get paid accordingly, and cost will inflate. And here we go again around the circle. All the plans will start down coding the patient complexity just like they do with service complexity.

And now we have another bot war. Meanwhile, premiums, are still going up. All of this reminded me of a conversation that I had had with Dr. Monica Lypson a while back, but as many conversations about heretofore at least, intractable problems go, it's still unfortunately, extremely relevant.

I wanna listen to this conversation with you and then we're gonna figure it out. And when I say we, I'm also including Dr. Siva in this because as aforementioned. This is a conversation that the two of us had, and I have definitely folded his insights into this mix.

Okay, so here we go. Monica Lypson, MD, MHPE. Welcome to Relentless Health Value. It's a pleasure.

Thanks for having me. Conventional wisdom would really presume that value-based care is good for underserved communities. Value-based care tends to address social determinants of health in ways that a fee-for-service system just simply doesn't. Do I have it right?

Yes. So as Einstein says, Intelligence is being able to hold two competing ideas in one's head at the same time and realize they're both true. So as a believer in value-based payment systems and a believer in that is probably our next best approach to the healthcare system, I do believe that that is a fiscal mechanism to improve the care of the underserved. Okay, so let's collect some wisdom here in our quest to solve the tension between a desire to maximize revenue and a desire to pay a fair price for value.

To define terms, when we say underserved here, consider that this could be a patient with preexisting conditions wherein those conditions are not controlled or resolved for many reasons, it could be someone with limited access to primary care or specialty care. It could be someone with social drivers of health given their health literacy or their income or employment status, or housing situation or language barriers. There's a lot rolled up into this underserved word is the point.

And if I have an incentive to up code, I could make the most of it. Or if I have an incentive to not serve a patient who might be hard to serve because I know I'm gonna lose money. Yeah, a lemon could have just gotten dropped. However, in the same breath if you have the wrong incentives, if you create perverse incentives, you actually might make known healthcare disparities worse by allowing for the fiscal system to pick and choose patients that are best suited to meet your quality metrics or to meet the demands value.

Ah ha, and this is what we mean by cherry picking and lemon dropping. Yes. I've even heard this applied in situations where patient shows up, has a BMI over a certain level, or had a heart attack previously. I mean, not necessarily even in an underserved community, and the cardiologist is like, Nope, sorry.

Can't fit you in. And I think that happens. If I wanna be Pollyannish, I will say, it probably isn't done volitionally, There's other things I can do to make me cherry pick those populations, right? If I only have office hours between nine and 12 on certain days of the week, potentially, I'm only picking either patients that are employed and can get off of work or patients that are retired, right?

Versus if I have evening hours, I'm much more Accessible. So there are ways that even the healthcare system that says, Hey, I didn't discriminate against that one person. There actually might be systematic and structural ways that the healthcare system might say to a certain group in the population, we're not really interested in taking care of you. Some have used this cherry picking slash lemon dropping as a reason why fee-for-service is better than value-based care.

What do you think of that position? I would say we've tried that experiment since 1965 and we have issues around disparities, et cetera, and part of that, the incentive to have a good outcome is not there. The incentive to have another visit is there. Fee-for-service incents, another office visit, and by the way, that another office visit also could mean a profitable readmission when the underlying problem or root cause of the issue isn't solved, like a readmission for heart failure.

But I also wanna point out something that Dr. Cristin Dickerson said in the episode with her about direct contracting for imaging. Link in the show notes. She said for Green Imaging, which is her company, fee-for-service works, the plan pays for only services rendered, and it's almost impossible for a directly contracted imaging contractor to refer themselves another office visit.

So the main perverse incentive of fee-for-service is mitigated. That's a point to ponder there. The only indictment I have on the fee-for-service system is that it's gotten us to where we are right now. That system is not working, right?

We have healthcare inflation. We have healthcare disparities. We are the nation in the, in the world that spends the highest amount on healthcare and our outcomes are middling, if not low, compared to our brethren. We haven't, there was no overt mechanism to, to actually measure the quality.

So again, if we even say most providers actually have the best intention for their patient, regardless if it's a fee-for-service or value based system, if you don't have systems in place to understand and connect the data points between the labs that might have happened outside of your system, the prior visit. If you don't have any connection in that system, even the provider trying to do best by the patient and provide a good outcome might be disconnected because the system is not in place to actually connect the lines for them or connect the dots.

Fair enough. Fee-for-service and most current network contracts available to self-insured employers and other plan sponsors, some exceptions aside, but most do not include really almost any robust quality assessment. Listen to like any of the episodes with Elizabeth Mitchell for more on this. Links in the show notes.

So for saying the word patient outcomes, obviously that's open to interpretation. What, what are the must haves for a, a value-based, patient focused fiscal system to create the outcomes that patients actually want and need and, and drive that clinician satisfaction? If I knew the answer to that, I could retire. Uh, here's, here's my perspective, The systems that have been shown to work are those who really take a, the VA is calling it a whole health model, or they're really looking at the whole patient as one.

So the patient is not a sum of its problem list or a healthcare problem. The patient is a person who presents often with a healthcare concern if they need healthcare. But those systems also typically offer access to things that make health a priority as well. So preventative services, a focus on both mental and physical parity issue.

And now what we're learning is we deal with the social needs of the patient and what the social context of the patient as well, and what that contributes to health. I can give you a real patient example. So I have a patient who is experiencing homelessness and has been without a permanent shelter for at least one to two years. And gets readmitted to the hospital.

It's been this interesting conversation because what will actually eventually save his life is giving him shelter. He has recurrent infections because it's gonna happen as long as he remains undomiciled. There's not a lack of knowledge that that is exactly what needs to happen. The hospital knows that.

Social worker knows that. I, as his primary care doctor, knows that. The hospital team that takes care of him every time and every time we think we get close to finding him shelter, there's some issue that we find he's not eligible. It's sad like it 'cause the easy fix is the cost of housing this patient for one year would be cheaper than his next admission to the hospital.

That's crazy. In a fee-for-service model like that patient, actually the hospital's rewarded every time he shows up with an infection. That's correct. Hold the phone.

Whole Person Health has entered the building. Before I go further, let me count how many others have said Whole Person Health on this show related to, for example, advanced primary care. We have had Dr. Scott Conard, Dr.

Grace Terrell. If you wanna go way back, Dr. Kenny Cole, Larry Bauer, listen to the podcast with Dr. Mick Connors, where we go off on this exact topic.

But here's why this is such a big deal relative to solving the tension between upcoding and down coding because none of those games are gonna be solved merely by getting clinical teams to take risk as aforementioned. But here's the thing, and I'm quoting Mark Weber right now, who was quoting Dr. Zane Gates In a recent LinkedIn post. "It's much cheaper to take care of people than it is to insure them."

Link in the show notes to that post. So bottom line, if we're paying for whole person health for that care in a really thoughtful way, many of the current upcoded profit centers determined to take advantage of the, how can we get paid the most to insure them games, these profit centers become cost centers. And look this by a long shot doesn't eliminate the potential for other Trouble with Tribbles , upcoding down coding problems, or the trying to constrain access or deny care to take advantage of the float problems, but I'd consider it a path forward that is at least better than the status quo at a minimum.

Or the funding streams don't line up, right? How can I take money from healthcare dollars and put it into housing? Our systems are set up such that it would be conceptually difficult to move that money around. How have you seen this work?

There's two examples I can think of. I've spent most of my career as a staff physician for the Veterans Health Administration. Actually, within that healthcare system, it was actually a lot easier to kind of order up housing, meaning there was an initiative over the last two decades to house to end homelessness among veterans. You actually had a much more nimble system where at least there were dollars allocated within that system to say we need to house her in a way that will take care of it.

Because we're actually probably the whole system will save money in some ways by that. But even in that system, those were separate funding streams. You saw this at the middle of the opioid epidemic in, uh, Vermont, where the hospitals in Burlington, what they sort of realized is if we could discharge our patients to a stable shelter and we'll pay for that healthcare system, maybe we can save money. There are a bunch of pretty good examples of whole persons getting taken care of out there, such as, of course, independent, advanced primary care practices.

But also the VA, the Veterans Administration, maybe also some of the really good FQHCs, Federally Qualified Health Centers. Nothing is of course perfect, but there's likely lessons to be learned in these examples. There's also community health plans that Dave Chase talks about quite frequently. Bryce Heinbaugh has pulled together some of these.

There's Primary Care For All Americans where towns are paying for its residents to get access to primary care and keep these people from becoming patients showing up in the ER. What's that quote? The future is here. It's just unequally distributed.

We're seeing this in small kind of pilot areas. But I think people are finally realizing we actually have to fix some of these structural issues if we really wanna fix health. We, we wanna be responsible stewards of that dollar. And I think the system should recognize that.

That's why you actually have to have patience in the room. That's actually why the community and the patient is probably the most important. Their bias is the bias we want in the room. This is where it all came together for me.

A patient's bias is actually the bias that we want in the room. And why do we need that many reasons. But here's one, because everyone else has bias. And the problem with that other bias is that the healthcare sector makes patient lives as the product, but nobody has agreed thus far on what guardrails that requires.

And again, there's nothing objective about a proclamation made by anyone who is well and truly biased. So why are we letting entities who we know are biased calculate, for example, their own RAF adjustments in the first place? I read an article a while back, I'll see if I can find it. That said, A more accurate way than any RAF is just use someone's zip code.

This is one of the most accurate predictors of risk that you'll find. Dr. Siva scratched his head at the moment 'cause we discussed this. He said, Yeah, what our goal should be is to find global call 'em ground truths about patients that cannot be gamed what we need here.

Are detached arbiters who are calculating patient risk and complexity. That should be the basis of upcoding or down coding. You can't ask biased parties to grade their own homework. The data has to be outside the control of those who clearly, again, have a very, very big incentive to put a finger on the scale.

Like credit scores, for example. There's plenty of industries do this. They have a weights and measures body that at least has some measure of independence and some measure of transparency. Dr.

Lypson, thank you so much for being on Relentless Health Value you today. My pleasure. Thank you for the opportunity. So look, I feel like I am beginning to cotton on to something that maybe should have been very obvious, but I really didn't see it before.

Here's the job to be done. Be good stewards of the dollar and do what we need to do to figure out how to give a competitive advantage to honest actors. Your next question will be, how do we do that? In situations where there is a risk adjustment framework or there is a sliding scale for bundles based on clinical complexity, get a neutral third party to risk score patients.

Just like we do with credit scores or in many other industries that have standards for weights and measures. Who is this? I don't know. Would I be kind of confident It's possible to do based on ground truths?

Yeah. And right now I'm thinking about that there's a case study from several years ago when some retail store, and this became a big brew haha. Knew that some teenager was pregnant before her parents knew the teenager was pregnant, buying habits, credit card swipes, copays paid, geofencing, right? So there's a lot of deep capacity at this juncture to do predictive analytics around individuals that has the potential to be more accurate than any chart review just saying.

Thank you so much to Dr. Siva. Thanks to Dr. Monica Lypson.

My name is Stacey Richter. This podcast is sponsored by Aventria Health Group, and I do also want to thank our 2026 series sponsor, Payerset. Thank you so much for the financial support to keep this show on the air. Hi, this is Dr.

Cristin Dickerson. Stacey, not only unearths and distills the problems with the existing healthcare system, she also offers hope for a better future for healthcare and for restoring trust in the system. If you rely on this podcast as much as I do, please make sure to follow the podcast on Apple Podcasts or Spotify and leave a review.