Improving Policy and Rural Healthcare: In Conversation with Brad Smith
I recently sat down with Brad Smith, founder of CareBridge, Main Street Health and Aspire and former head of the Center for Medicare and Medicaid Innovation (CMMI), to discuss the novel ways he’s bringing risk-based models to rural healthcare and Medicaid and hear why he thinks CMMI programs haven’t worked to date. Brad’s takes were really thought-provoking and you can listen to our conversation on Spotify or Apple Podcasts.
Our conversation was a continuation of my exploration of the current state of value-based care in this piece and discussions with Adam Boehler, Mandy Cohen and Tom Lee.
Brad has a fascinating perspective to add. He was the co-founder and CEO of Aspire, the largest home-based provider of palliative care, which focuses on optimizing quality of life and reducing suffering for people with serious illness. Brad ultimately sold the company to Anthem and then had a stint in government. He led CMMI and had roles with the Domestic Policy Council and Department of Health and Human Services, playing a key role in the government's COVID response. Now, Brad runs Russell Street Ventures where he's focused on launching and scaling tech enabled businesses in value-based care. His first two businesses are Main Street Health, focused on rural seniors and CareBridge, a healthcare unicorn focused on Medicaid patients who need assistance to live independently.
A few things stuck out to me from my conversation with Brad:
1. Expanding risk to rural and Medicaid populations is difficult but has unique advantages.
Most of the early risk-based primary care companies were focused on Medicare Advantage patients in more urban areas. Medicare Advantage populations generally have higher reimbursement rates than Medicaid and require a more similar care model given most older people have health issues. It’s easier to make the unit economics of these businesses work when partnering with larger urban practices given the ability to share resources between them. In expanding these models to rural areas and Medicaid, Brad faces challenges. But interestingly, focusing on these populations provides some unique advantages.
Given that it’s relatively easier and more lucrative to work with urban practices, many companies seek to partner with them. This means a practice may have different partnerships for its Medicare Advantage risk arrangements, clinical trials and revenue cycle. It is objectively harder to make the economics work to partner with rural practices. But if Main Street Health can make that work they may be able to offer a much broader set of services which can help cover the relatively higher cost of supporting each practice.
In Medicaid, Brad has seen a really interesting dynamic with CareBridge. Medicaid Managed Care Organizations (private payers) apply for a limited set of state contracts. They are now using CareBridge as a way to stand out, or keep up with, their competitors. This same dynamic doesn’t exist in Medicare Advantage where there isn’t a limited set of spots for private payers. It provides a strong different sales value proposition.
2. The voluntary nature of CMMI programs makes generating savings challenging.
Brad raises some interesting points around the mixed success of CMMI programs. While running CMMI he did an analysis of the 54 programs the agency had run over the past decade. A particularly poignant point he made was that setting a voluntary program on average benchmarks generally just leads only people who are already performing better than the benchmarks to apply.
3. State by state differences in Medicaid make building product hard.
The difficulties in building a Medicaid business because of state by state differences has come up a few times in previous conversations. Brad outlined a similar challenge in building CareBridge: there's about 127 different assessments that states use to determine how many hours or types of care people get in Long Term Support and Services. Building product to meet that is challenging.
4. One strong process for starting healthcare companies is looking for models applied at a smaller scale.
Brad outlined a process for starting Aspire that was reminiscent of what Adam Boehler had mentioned about Landmark. Adam had said
“The United States is a big place - usually you can find places where there’s data from things that have been applied.. there’s always analogues to our businesses. In the case of Landmark, the Department of Veterans Affairs (VA) published a couple studies on their home-based primary care model...we read the studies and thought we could unlock a lot of value through doing it.”
When Brad was thinking about starting Aspire, he similarly found a bunch of small nonprofit hospices that had achieved impressive outcomes on grants but then shut down their programs.
This company-building approach of looking for models applied at small scale is quite compelling in healthcare.
Check out a full transcript of our conversation below:
To kick things off, you've had this really fascinating career arc. I think when you were in college, you were thinking about being a doctor, you then entered the policy world working in education, and then you ultimately arrived in health tech with founding Aspire. I'd love to hear a bit about that journey and how you ultimately came to Aspire.
Yeah, it's been an interesting path, and probably not one that I could have ever predicted. I've always had a passion around helping people and around a mix of nonprofits and government and honestly, business wasn't something that was at the top of my list after college. My first job out of college was working for a campaign and I drove a guy who was running for the US Senate around Tennessee to all 95 counties three times. And it was a great experience and I got to learn just a ton. You go out you meet farmers, you meet doctors, you meet everybody in between and after that I ended up going to grad school and when I moved back from grad school, the guy I had worked for, his name's Bob Corker, ended up being in the Senate for a while and introduced me to Bill Frist. At the time, Bill Frist was getting really interested in education policy. That had actually been what I studied in undergrad and grad school. And we were working on what we thought was going to be a book about education policy, and specifically K-12 education policy.
We never ended up writing the book, but we ended up launching a nonprofit. So I ran this education nonprofit in Tennessee, I went and worked for the governor doing economic development. And I was probably in my early 30s, and trying to figure out what I was really going to do. And I was back one Christmas and my brother who was in med school at Harvard took a year off to work with Atul Gawande on research for his book Being Mortal. And he just started talking to me about the work they were doing around palliative care.
I live in Nashville, my family's in Knoxville, about three hours away, probably three or four months later, my grandma's started getting sick, she ended up going in and out of the hospital, I was driving over to see her, she ended up passing away without ever going to hospice. And so I had this very personal experience with it. And my brother had talked to me about it. And so I actually went back to Bill Frist who I hadn't actually talked to in a couple of years and said, "Hey, I have this crazy idea to start a palliative care company." And he was kind enough to support me and back it and that's how we got kick started on Aspire.
When you started Aspire, I imagine the conversation around palliative care was pretty different than it is today. As you've been in the space for the last decade how has that conversation changed with physicians and patients? How has palliative care evolved and where is it headed?
Yeah, well when we were starting Aspire and going out and talking to investors they were really worried about the perception. As you may remember, it had been a very political issue for a short period of time 10 years ago.
This was the whole death panel part of the Affordable Care Act discussion?
Exactly. And so there was a big concern from investors about would health plans really do this etc. But what we had seen across the country is that there were a number of small nonprofit hospices who were setting up these programs very successfully. And typically, they would set them up for a year or two, they would publish a paper, they would have a grant then the grant would go away and the program would go away, but the clinical model was really working. And so what we set out to do is really show that it could work at scale. And I think in many ways, we did that with Aspire. We grew to about 25 states, we served over 100,000 patients before we ended up selling it, and Anthem's done a good job continuing that on. But I really think between what we did and what a couple of others in the space did, who were either doing palliative care or adjacent complex home based visits, it really showed that you could deliver this care in the home in a way that would be really helpful to patients. And it could help lower cost.
And I think you've now seen CMMI and more payers fund these initiatives. You've really seen this big movement over time. I actually saw a chart, the other day, that said in traditional Medicare home based visits have gone up slightly, but in Medicare Advantage, they've gone up about four times, right? And so I think you're seeing this innovation that people hoped would occur in MA, actually occur there.
I think as the space matures there's also a lot of learnings happening. I think that which patients it makes the most sense to deliver these programs to is a really important question. Sending a nurse practitioner and interdisciplinary care team to the home is not cheap. You have to figure out how you target it to the right patients, and for what length of time. So I think this space is evolving in terms of how to titrate for lack of a better word, that intervention. You've got one time folks like Dispatch and Ready Responders, and then you have longitudinal things like Aspire and Landmark and everything in the middle. And I think it'll continue to refine itself over the next five to 10 years.
Switching over to some of your work at Russell Street Ventures today, you’re doing some really interesting work with value based businesses. If I think about the space at large, it seems like a lot of the early value based businesses were focused on general primary care models largely for Medicare Advantage where there's a consistent care model given members are old and generally have some burden of illness. Most companies have been in relatively dense areas. At Russell Street, you’ve expanded where these risk based models are focused, targeting rural areas, more specific populations and Medicaid. I'm sure you do a ton of analysis before spinning up and starting these businesses. How do you think about the populations where it does and doesn't make sense to build some of these risk based models?
Yeah, I'd say it's a mix of both qualitative and quantitative, right? So in many ways, some of the topics that we've gotten interested in like Long-Term Support Services (LTSS) where patients are getting typically a non clinical caregiver in the house, those were patients that we bumped up against at Aspire. In rural, those were areas I've been. But we definitely do a lot of analysis, And so, as we were thinking about where we wanted to build other businesses after Aspire, we downloaded data around where the Medicare expense was, where the Medicaid expense was, and where the commercial expense was. And we systematically went through and said, where are there big pockets of expense, where we don't think there's a lot of solutions.
I think some of the best advice I ever got about picking where to start a business was from Annie Lamont, who's been one of our key investors at Oak. Her advice was, "Hey, it's way easier to build something really special, if you're the first person starting in a space versus if you're the 10th person starting in a space. And honestly, you also can make a few more mistakes and bumps along the way and still be okay." We've really screened that as a core criteria for the businesses that we've started. We want to be first or second in whatever space we're going into.
Starting with Main Street, which as you've alluded to focuses on a rural population. Our listeners are probably familiar with this first generation of predominantly urban primary care providers taking on risks like Oak Street, ChenMed and Iora. As you thought about starting Main Street and building that model I'm curious how you thought about what changes were required to those traditional models to make things work in this specific setting?
So it was actually an interesting journey, working on the Main Street idea, because we ended up with something that looks like a rural version of what the urban MA guys are doing, which I'll come back to, but we actually ran down a lot of other ideas before we got to that.
Really, what other ideas?
Our first idea was trying to partner with rural hospitals because we read about all these hospital closures. And in many of these communities, they're the largest provider of care. And what was interesting is when we went out and spent a lot of time with them, we realized how hard of a financial set of circumstances they were in, and also honestly, how much capital investment was required in their physical plants, for which was typically disproportionate to the level of care they were either providing or could provide in the future. And so it really felt like there was this capital constraint.
The next thing we looked at and thought really hard about was urgent care providers and starting with some of these, if you've been to some of these small communities, almost all of them actually have an urgent care. And we actually got in a really deep diligence on one and I think what we realized around that is putting a new box in one of these small communities, it's just really hard to make the volume work right? Another idea we ran down was there's a lot of independent pharmacies left in these rural communities. And we actually piloted a model trying to do things in the independent pharmacies and what we found is while people may be getting more used to going into CVS or Walgreens that have created these clinics inside, these independent pharmacies haven't; they look more like a retail store and you're putting up literally like a table in the back corner. And you’ve got to do the build outs that CVS and Walgreens are doing to really make that work. And so really, the way that we ended up back at what looks a lot more like a Medicare Advantage MSO risk model, is by running down a lot of other ideas that it turned out, at least for now, not to be the right idea for how to get started. And then we did look at what a lot of the urban folks were doing and said, "Hey, look there's a model where people are owning clinics,” which we didn't think made sense in rural areas, just given the volume issues.
There's a model where people are partnering with Primary Care Physicians (PCPs). And what we knew is in rural America the PCPs that are in these small towns, many of them went to medical school in larger cities and they inherently chose for some specific set of reasons to move back to their local community. And in many cases, they're the leader of the town or one of the key leaders. They run one of the largest businesses, they're a deacon at church, they run the Kiwanis Club. And so being able to partner with them, to help them build out their practice and serve more members and do it in a more thoughtful way was a really unique opportunity. And then when we thought about what population to serve, we actually started trying to build a Medicaid model. As we got into the numbers, it just became really hard to see how you could do risk based primary care in Medicaid in a financially viable way. And so we decided to do what a lot of other people are doing, which is launching Medicare Advantage. There's a lot of reasons that makes sense. And now we're moving into traditional Medicare with an eye for how we might move into commercial and potentially Medicaid with some different strategies. But I think we ended up probably a lot closer to what the urban guys are doing than we thought we would.
There's a lot of differences in how you have to run these small practices. So one on the challenge, and one on the opportunity. On the challenge side, the practices are small. They have one doctor, maybe two doctors, if they're big, one or two Nurse Practitioners (NPs), 10 or 15 employees, and asking them to do more with the people they have is really tough. And so I think our best innovation, for lack of a better word, has been putting this health navigator is what we call it. It's like a community health worker who makes $40,000 or $50,000, which is a great job in a lot of these rural communities, and putting them in the clinic to be that extra set of hands. And that's made a huge difference. And I think the opportunity that we have is in some of these large practices that Agilon and others partner with, they're using somebody for their MA risk, but they have another partner for clinical trials and another partner for their RCM and another partner for whatever. And what we're finding in these small practices outside of pharma reps, nobody's going to see them. So the ability to serve them and provide them a lot of different services across a lot of different populations we think presents a different opportunity than if you were in one of these more urban environments, where you can't be everything to somebody, you’ve got to be a really specific version of something. So we're excited about how that might evolve over time.
To your point, there's probably so much you could do especially having someone on the ground at the practice. How do you think about drawing the lines of saying, "Okay here's what we'll do for now, in phase one of the business, here's what we'll do long term and look, this is never going to be in scope." Because I'm sure there's no shortage of directions, you get dragged in?
Yeah, it's a great and highly relevant question. We built a long list of ideas, as you can imagine, for all the things you could do. And we're kicking off this six month process. We built a little team of former consultants, bankers, all that stuff, to really think about where else could we take the business? And I think as we do that we're thinking about three screens. One screen is, is it going to be economically sustainable over time? Because obviously, we want to make sure that whatever we do, we can sustain and scale. Number two is, what is it that our practices really want. So we're doing surveys of our practices with a list of different options to see what they really want. And the third is what operationally fits within the other things that we're already doing.
And so as we put things through that screen we think we'll take our list of 15 or 20 ideas and get it down to two or three, at least for the immediate future. And then we'll create this pipeline that as we grow and scale, we can think about doing more things. I don't know if there's anything that we're going to carve out of scope for forever from the start. I think I could see myself, not a commitment, but I could see myself building this business for 10 or more years, right? And so in that world, putting everything in scope sounds really fun and interesting so we'll see, but we're at least trying to use those filters for figuring out where to get started.
So today what are those folks on the ground mostly doing in partnership with these primary care doctors?
They're doing the things around value based care that needs to happen in a practice. I'll walk through a couple. So one is there making sure patients come in. It turns out not all patients come in at the right visit cadence. And so they're helping the person at the front desk reach out to patients and make sure they're coming in. We focus a lot on doing annual wellness visits, we call them comprehensive care assessments with patients when they come in. So obviously, the physicians and the nurse practitioners do that. But our health navigators support that. We also do what we call these check in visits, when folks come in, if you think about that time between when you get pulled back to the patient room, and the doctor comes in that's when our health navigator will pop in and see you.
We're getting ADT feeds in all the markets. So when somebody goes to the hospital, we'll call them within 24 or 48 hours and schedule them to come into the PCP. We're helping close quality gaps, we're doing a high risk case management program. So really the health navigators are taking this population of patients and doing the things that the doctor wants to have happen, but just doesn't have the time or the extra set of hands to do and being that extra set of hands.
To what extent is virtual care part of the model? I feel like the easy answer to making the model work with lower density populations is doing more of that. I'm sure you've explored it, to what extent is that part of the model today?
Yeah, so today, it's not a big part of the model. Actually, interestingly, we built this app that we thought people would use and nobody ended up wanting to use it to talk to us. So a good learning. But it's going to be. When I talk about that strategy project, one of the things that's very high on our list is specialty care. And we have some ideas around how you build a layered model from econsults to virtual telehealth to contracted network and step people up through it in a thoughtful way. And so I definitely think as we move forward and get more sophisticated, figuring out when and how to embed virtual care for the right set of patients, especially around specialty services, is something we're going to be focused on. But honestly, we started old school getting people to come in and see their PCP and meet somebody face to face. We definitely plan to weave in the virtual as we scale up.
That specialist work seems particularly impactful. I imagine in a lot of these rural communities the access gaps that exist are probably even more pronounced in terms of wait times to get into see specialists and just general availability. So the econsult and virtual approach makes a ton of sense.
One of the interesting things we've seen about rural communities is, only about 30% of the specialty care typically either happens in that community or the neighboring communities. About 70% of that is going out to a larger city. And so we really feel like making sure it's the right specialty care that's getting referred out, and that they're getting referred to the right specialist, creates a big opportunity.
Early on when you were doing some of the basic evaluation of this business, what gave you the confidence that the economics would work? Now I'm sure you're starting to see some of the early results but it's more challenging to make this work with lower patient volumes. How did you get over the line and say, "Yep, I think there's actually a business here that will work from a bottom line perspective?"
Honestly what's helpful about having these urban folks public is you can go online and download their investor reports and see their economics. So I think you start from what's publicly available. And obviously, being in other places I've been, you have some sense of that. And then you say, okay, "Hey, what has to be different about our rural setting?" So for example, we thought we needed these navigators, which is a cost that not everybody else incurs, although they incur other costs different ways, and then you build can you support that. I think the reason we started in Medicare Advantage versus say, Medicaid, where we actually wanted to start initially, was the economic picture became more clear, and we really believed we could use MA, for lack of a better word, to help fund our navigator that we're putting in the clinic. And then as we expand into other populations, where we have to have a more efficient operating model, we can take that fixed cost navigator and spread it across a larger population.
I'm sure the primary care doctor really appreciate having that person support a broader patient panel. Switching gears a bit to Medicaid your second business CareBridge is focused on the LTSS space. For listeners’ background LTSS is Long Term Services and Support. It’s a Medicaid program that helps people with disabilities and chronic illness live where they want. It includes a variety of home and community based services (HCBS) to help with daily living like eating and bathing. How did you get interested in the LTSS space and managing the medical and community services spend for these populations and where you saw the opportunity for CareBridge?
I’ll make a couple of comments about Medicaid. So I think a lot of folks want to serve the Medicaid population. And I think when you think about who is the Medicaid population, there's a bunch of things that are pretty different from Medicare. Number one is the Medicaid population is not a cohesive population. There are different waivers, there are different ways you qualify etc. The diversity of a mom who's pregnant and somebody on LTSS is massive, right? And so Medicaid is really a bunch of different populations rolled into a single government program. You have to figure out which population you're going to serve. There's a ton of state to state variation, Unlike Medicare Advantage, or Medicare, where there's one national program, there's 50 Medicaid programs.
How has the state to state variation impacted your business?
It has a huge impact. So in our business where we're doing home and community based services and managing them, there are different waivers in each state, meaning different populations that they serve. So in some states they serve Intellectual and Developmental Disabilities (IDD), in some states they serve brain injury, in some states it’s more like aged and disabled. There's about 127 different assessments that states use to determine how many hours or types of care people get.
Wow.
Yeah, it's crazy. They all have different billing codes. So every state has, not every state, but there's a huge variation by state and billing codes. Some of them use these things called modifiers, for each code that changes the rates. Other states have modifiers at the patient level, not at the code level, it's this whole crazy amalgamation of things. And so it has a huge effect on how you build the business.
When you build the business, you have to build it in a modular way. And so even though there's like 127 assessments, they're not completely different from each other, they're just slightly different. The way you build your assessment is you end up building a really long assessment that's inclusive of all the questions everybody asked, but then by state, you're able to pull out sections 1, 3, 7, and 12, to ask a patient. So you're both able to standardize and to specify the state.
And I think knowing that that's what you have to do in Medicaid, we started building it that way. But you don't build a Medicare business like that, they pretty much work in a single way. And we just knew we were going to have to have different delivery models. And so figuring out where our infrastructure had to be adjusted, really, from the start, honestly, to be able to accommodate that was really important.
Putting your policy hat on, is that state by state variation good? I can see arguments on both sides. That variation does allow a lot of states to experiment with new things, and you could generate evidence early on what works and what doesn't. At the same time as an entrepreneur, it's tough to build with 50 different ways of doing things. In terms of that balancing act of encouraging state by state innovation, and then making things standard to make it easier to build businesses, where are we? What policy changes would make it easier to build in Medicaid?
I think innovation is good and I think the more that we can have parties, besides the federal government, who can drive that innovation, whether it's states or Medicaid Managed Care Organizations (MCOs) or MA plans, in my mind, that's where a lot of the successful innovation is coming from. I ran the CMS Medicare Innovation part for a while and long story short is we've spent $15 billion, and it’s basically not working.
You want that innovation at the state level and in the private sector. But I think if you're an entrepreneur, you have to enter it like anything that's more complex with eyes wide open. As we've shown with CareBridge, we built a really good business really quickly honestly. We did it by really understanding how Medicaid was different from Medicare and building everything from our team, we've hired a ton of folks who've worked in state Medicaid agencies for example, to how we built the IT infrastructure, to how we built the operations, just all with that differentiation in mind.
So I think it's a little bit more complex, but it's not hard if you know you're doing it and I think it'd be more hard if you assume one model is going to work everywhere. And then you're three years in and you realize that's not the case. But if you start with that in mind, I think it can be good. And the other thing I’ve found, and this is more of a business thing that we've been surprised by at CareBridge, it's really the reason CareBridge is growing so fast, is the RFP process in Medicaid, creates this real competitive dynamic between the plans. And so what we're seeing, and this is driving a lot of the CareBridge growth, is after we're in a market and successful with one or two MCOs we become the thing that every MCO needs to be competitive in their RFP. And so we're finding several states where literally, we're in every RFP that's getting put into the state Medicaid agency.
And I think we didn't realize how powerful that sales mechanism could be because there's just nothing comparable and the plans don't need you for their MA bids to CMS. But because the Medicaid bids are more competitive and not everybody's chosen they really try to get a leg up and that creates a big opportunity and an incentive honestly, for them to innovate and try to do new things.
That’s really interesting on Medicaid Managed Care Organizations trying to differentiate their bids. What is the gap in the LTSS space? Plans are clearly flocking to your solution. What were you looking at, at a 30,000 foot view that got you excited about the CareBridge opportunity?
What got me interested initially was, at Aspire, we had home and community based service patients that we served, and they would have these paid non clinical folks in the home, and we always knew we should leverage them, but we never really figured out how, because they were 5% of our patient population. And so I always had this idea that, "Wow, if we could have gone back and built around that caregiver in the home what could we have done?"
And so really, that's what we set out to do is really serve these patients who are receiving home and community services, have a paid caregiver, or sometimes an unpaid caregiver family member, most of the time a paid caregiver in the home, and how could we both work with that caregiver to integrate a telehealth virtual clinical team? Because you already have that caregiver face to face. And at the same time, how can you help manage the home and community based service side of the spend. Because typically, these patients, they're about 2% of the Medicaid population, but they're about 20% of the total Medicaid spend.
The average spend is about $50,000, half of it is on the medical side, and half of it is on the home and community based service side. So you have to help manage both parts of it. And I think what we found is that because there's 127 assessments, and because there's 50 different states, there's almost no common standards across the space. And these programs are really created to make people more independent right? The whole point of having HCBS services was to keep people out of nursing homes. And I think what's happened unintentionally over time, is really, the programs have made people more dependent on them.
I'll give you an example that we see that's really common. Let's say a patient is having a hard time taking a shower. The way most of the systems work today is they're going to pay for more caregiver hours for somebody to come give you a shower. But what happens is, when somebody does it for you over time, you become more dependent on them and less independent, right? An alternative, and we use occupational therapists and physical therapists, they'll be like, "Hey, what if we get a grab bar in your shower? What if we get you a chair so you can sit safely holding the grab bar, and what if we give you four weeks of OT or PT who can really train you and come watch you and make sure you can do it safely, right?"
That doesn't work for everybody. But there's a subset of patients for whom that works. And when they're now using their muscles to get in and out of the shower, they actually don't deteriorate in the same way as if somebody was rolling them into the shower or bathing them and rolling them out, right? And so I think there's really this opportunity to really help people be more independent for longer and lower costs at the same time, especially when you can wrap the nurse practitioners and social workers and all that around that non clinical person in the home. And I think that's what we've seen and what we're proving out at CareBridge.
It feels like to make that work, you need a payment model where you're responsible for the cost of care across both medical spend and at home care spend as well, to be incentivized and have a model that works when you make that change, versus just sending in the caregiver for more hours. It seems like, correct me if I'm wrong, but we’re in the early days of understanding the efficacy of some of these interventions for subpopulations in at home services. So how do you think about the state of evidence generation? Do we know where sending in a PT and OT for this patient really does actually help, whereas, maybe it's less impactful here? To your earlier point it seems so much of the focus for these businesses now is figuring out okay, what's the specific populations where at home care interventions make sense.
I actually think we're in the very early days of value based care in its entirety as a country. When I was at CMS they had me go through and look at all the work that had happened over the past decade around CMMI. And we had launched about 54 different models. And I'd say, if you think about three things we were hoping to change, we had various different levels of impact around those things. So one thing we were hoping to change was how people got paid and that part was working. We had a million providers who were in CMMI models and people were getting paid differently. Now I think what people believed getting paid differently was going to do was lower cost and increase quality.
And we saw some signs, but not very strong signs of increased quality in certain programs, but they were definitely not overwhelming in any way. And on the whole, the programs were losing money. We had four or five that were saving money at some level, two at substantial levels. But we had several that were losing honestly, billions of dollars. And so I think when you look in aggregate, where are we on value based care? Yes, we figured how to change how we pay people, that's definitely happened. Is quality going up? I don't know. But people are definitely getting more care coordination for sure. And is cost coming down? I think that's a very hard thing to answer yes to across the board.
So I think, whether it was in my CMMI role, or as we're building businesses, I start from the assumption that it's very hard for these things to save money, and that you have to have a very specific focus on a patient population and very clear value levers. And I think people probably five or 10 years ago had the idea of, "Hey, if I have somebody call you more or come to your house more that inherently is going to result in more savings." I just don't think that's true, especially when you think about it on a net basis, that includes the cost of the intervention. And so I think we're going to get more refined and sophisticated in all these models, both who we focus them on, what the intervention needs to be, and how long it needs to be, for it to really both generate higher quality for the patients, but also honestly, a lower cost for the system. And I think we're in the very early days around that still.
Obviously, it's a sobering analysis you did at CMMI looking at these previous payment models. I totally hear you on the need to get more specific in the design of them and you're obviously living that out in the businesses you're building. From a policy perspective, what lessons do you draw from the fact that about $15 billion have gone into these models, as you think about the role of CMMI going forward?
I think there has to be a real question: what's the goal of CMMI? And I always thought one of the goals, not the only goal, but one of the goals was to lower cost. And I think the takeaway from looking back at the last 10 years is, it's not doing that effectively. And the question is, why is that? There's a lot of complicated reasons, but the core driver is, most of the models are voluntary. And so to get people to participate in a voluntary model, I'm oversimplifying, but you basically have to pay them more. And CMS came up with a bunch of sophisticated not sounding like price increases ways to pay people more, whether that was upfront PMPM payments, or setting favorable benchmarks and letting you select the people below the average to put into the program and all these other things that were well intentioned and intended to drive participation. But the result was, they didn't save money.
And I think to the extent that lowering health care cost is a goal that we want to pursue, you've got to make really hard decisions. And I actually went to Seema Varma [CMS Administrator] and Alex Azar [Secretary of Health and Human Services] a couple of months in after doing this analysis, and I said, "I'm not saying CMMI can't work, but I'm saying it's very difficult for it to work, because whoever's running it for five or 10 years, if you want to save money, they have to make hard decisions that go against special interests over and over and over and over again. And they're not going to be increasing rates, they're going to be cutting rates in aggregate, right?" And, yes, there are efficient ways that you can do that, but most of the ways people were making money in the programs was honestly, arbitrage of different kinds.
And so I think that's just the facts and the reality, we can all decide what we want to do about it. Maybe we don't care if the programs save money, maybe we do, there's different outcomes, but I think the facts are true. And so I think there's going to be a lot of soul searching, I think, at CMMI and CMS about if we're really trying to lower healthcare costs in the country, how are we going to do that? And the thing I say to everybody is, I had lots of people come pitch me ideas at CMS, but every one of them was a version of will you pay me more? They sounded lots of different ways, but they were all about people asking for more money, right? And it is hard to sit in those roles and say no, but we need people in those roles who will do that, because it's to everybody's benefit to do that, but in the moment, it's really hard when you have special interests pushing you on things.
If you really wanted to reduce costs, what are some of those difficult decisions you'd make? I think you alluded to part of it maybe being requiring folks to shift over to some of these models versus making them voluntary. What else is on the list?
I’ll start with CMMI and then go a little bit broader, I think in terms of CMMI, a lot of it does have to do with not making the models voluntary and/or learning from some of the things that we figured out over time, which I'm not saying we knew 10 years ago when we started, but we know now. So as an example, if you set a regional benchmark or an average target price, and you let people voluntarily pick, the nature of an average is some people are above it and some people are below it. So if you get to pick, which ones do you think go into the program? The people below it right? And so we shouldn't set an average target price. Or if we do, you’ve got to be really careful about how you do it. We always had folks come to us and say, "Hey, we need all these upfront payments to provide care coordination." It’s really true in the oncology model for example.
And so we made these big payments, and what turned out is, they lowered medical costs, but it was only 10 or 20% of the upfront payment, we paid them. So you can pay people upfront, you’ve just got to put them in downside risk if you pay them upfront. I think there's some technical things you can do that could really help with the CMMI models, but none of them are easy decisions.
You’ve got to have really a lot of gumption honestly, to sit in that role, and continually make the decisions that are hard, which is why I told Seema and Alex I thought I was going to be a hard road for CMMI to really save money, not because it wasn't possible, but because it wasn't going to be comfortable for the person sitting in that seat if they made those decisions over time.
I think there's things you can do outside of CMMI. Generally, it's really hard for the government to cut costs. If you just look at history, social programs don't go away, they only get bigger, basically. And so I think you've got to be really creative around how you create incentives around those things. There are things you can do around competitive programs and competitive bidding and these kind of things that really create market incentives where the government can get what they want, but doesn't have to be the bad guy directly cutting the costs and things like that. And I think there's a lot of innovation opportunities to really allow people to, in a competitive environment, compete on quality for sure, but also on price. I think that can help lower the overall cost. But those are pretty big changes.
What's an example of what that might look like?
So one example is something that we proposed at the end of my work at CMMI that got rolled back, and it was called the direct contracting geographic model. Basically the idea was that you would take a couple MSAs, and you would basically have providers and plans bid to do care coordination for the traditional Medicare patients. And it was set up in a way that patients would get all the same benefits they get today, or better benefits, they would still have access to all the providers in Medicare, and all the providers would get at least 100% of Medicare. So we thought through all the downsides and corrected for them.
But what it was going to do is by only having a couple of folks, call it two to five folks in a region, manage these lives in a competitive process, we had talked to enough folks that we knew they were going to be aggressive on price. We had heard this, it never got bid out because I left before it happened, but we had heard people were going to bid five to 10% below the Medicare prices on the business. And so it was upfront savings. There are ways to do these things, they're just complicated. And I think the thing that was a bummer to me, is that rather than looking at some of these more complicated things and tweaking them honestly, the CMMI folks rolled back everything we did in the last year I was there.
And I think that that doesn't set up CMMI for long term success. I'm not saying any idea we had was perfect or anything like that. But I think if you end up in this world, where... CMMI was going to be a hard battle no matter what. And now, if you're in this world, where it goes from one administration to the other, and they change policies, people can't participate in the programs reliably, etc. And so I think there's been a lot of risk created. It's hard for me to see a path where CMMI is really successful long term.
Fascinating, I feel I could talk to you all day about this, but I want to be respectful of your time. My last question would just be, you're building some really interesting stuff at Russell Street Ventures, what's next for you guys?
Yeah, well, I think when I came back from DC, my hope was to build three to four companies. I think, what we've realized as we've got these two going is, it reminds me how hard it is to build and scale companies. So we've got one other idea we're baking on and we're going to see if we launch it later this year or if we stick with the two that we have, but I think we're going to stick to a really small number of companies and just try to do a great job doing two or three things. And I think if we pick populations that are both big opportunities, but also where there's a big need I think that will keep us busy for the next couple of years.
Well Brad, thank you so much for taking the time.