Farzad Mostashari, M.D., on the Future of the MSSP—and Its Challenges
This spring, Farzad Mostashari, M.D., ScM, and Geoffrey Moore, Ph.D., authored an article in The New England Journal of Medicine entitled “Crossing the Chasm: How to Expand Adoption of Value-Based Care.” Dr. Mostashari is the former National Coordinator for Health IT and is the co-founder and CEO of Aledade, the Bethesda, Md.-based physician enablement company. Dr. Moore is managing director of Geoffrey Moore Consulting, South San Francisco, Calif., and chair emeritus of The Chasm Group, South San Francisco, Calif., and the author of Crossing the Chasm: Marketing and Selling High-Tech Products to Mainstream Customers, published in 1991 and revised in 1999 and 2004.
In the article, published online on March 19, Drs. Mostashari and Moore wrote that, “Beginning in 2005, with just 10 practices and about 150,000 beneficiaries in the Physician Group Practice Demonstration, participation in accountable care grew rapidly with the launch of the Medicare Shared Savings Program (MSSP) in 2012, to over 10 million beneficiaries by 2018. New accountable care organization (ACO) models launched by the Center for Medicare & Medicaid Innovation (CMMI) in 2019 provided another burst of growth in alternative payment models, but it appears that the accountable care effort has slowed since then. In 2024, the CMS announced that approximately 13.7 million people with traditional Medicare were aligned to an ACO, up 3% on 2023,3 but that is fewer than half of the approximately 29.7 million people enrolled in traditional Medicare, and far short of the pace needed to meet the goals on moving payment and beneficiaries to these new models. However, there is no going back on accountability. Combined spending on Medicare Parts A, B, and D rose from US$583 billion in 2013 to US$1 trillion in 2023, and is expected to double to US$2.1 trillion by 2033. When Congress inevitably looks for cost savings, a pivot away from value-based care would mean choosing between higher taxes, reduced benefits and slashing payments to providers.”
What’s happening? “Progress is stymied because the things that attract early adopters and innovators are not the things that will entice remaining providers to make the leap from traditional fee-for-service payment models to value-based care,” Mostashari and Moore wrote. “There is a chasm to be crossed as value-based care moves from being something that a relatively small number of forward-looking provider entities embrace to a default model that is widely adopted, thus becoming the new mainstream.”
Fundamentally, the authors noted, early messaging led to an initial surge in participation in the MSSP. But, they wrote, “Aledade has recruited more than 1,900 practices and health centers into its network (through more than 200 value-based contracts spanning 45 states and Washington, DC, with more than 2.5 million patients’ lives under management), but we have also encountered thousands of other practice decision-makers who have declined to join. We see that the concerns and priorities of practices that have resisted are substantially different from those of the early adopters and innovators who fueled the early growth of the program.”
Mostashari and Moore argued that the key to ongoing progress will be to encourage forward those physician practices with the most to gain. “For accountable care, the segment of health care that is most underserved by existing payment models is primary care,” they noted. “Primary care has the most to gain because fee-for-service medicine has not worked for it, and because its incentives for reducing cost of care are not conflicted. We believe this explains why primary care–led ACOs have generated the best performance (i.e., greater savings for the Medicare program, in terms of both gross and net savings) in the MSSP, Medicare's largest accountable care model.8 However, we still have only about 45% of primary care providers participating in Medicare ACOs,9 so the focus needs to be maintained on the primary care market until the chasm is crossed and the new value-based mainstream is created. Because the MSSP has made inroads in the primary care community, we suggest that the MSSP remain focused on primary care and serve as the chassis for CMMI models that seek to integrate specialty care into ACOs.”
In other words, the authors argued, CMMI must focus on models that seem to offer a clear return on investment for the physician practice leaders willing to consider reengineering their practices for greater efficiency and effectiveness, and to improve patient outcomes. The era of early experimentation is over, they wrote, and they strongly urged CMS officials to rework its benchmark system for performance and to address the inevitable cash-flow challenges for groups participating in the MSSP.
“The new presidential administration will be the one to see this generational effort through to 2029. We can get to near-universal adoption of value-based care in Medicare with the way new technologies bridge the adoption chasm — in this case, by focusing on a niche market (go deeper with primary care); defining the competition (fee-for-service is bad — for patients, practices, and society); switching from bells and whistles to proven solutions (MSSP is the chassis); assembling the whole product (cover the whole patient panel); and making space for the simplifiers (tap into physician enablers). Put simply, what primary care practices fear in transitioning to value-based care must be made less onerous than what they should fear from the fee-for-service status quo. Because if we fail to cross this chasm, to create a new value-based mainstream, we risk a bleak future in which our health care system continues to deliver substandard care at a high cost, and primary care continues to atrophy. Our aging population needs and deserves a better future,” they wrote.
Per all of this, Dr. Mostashari spoke recently with Healthcare Innovation Editor-in-Chief Mark Hagland regarding what needs to be done going forward, and the prospects for change at a time of uncertainty. Below are excerpts from their interview.
If you had to give an “elevator speech” about your and Dr. Moore’s article, how would you explain things broadly?
I think what really struck me was that I was rereading a classic book on technology adoption that Geoffrey Moore wrote, Crossing the Chasm. And what he describes in the book is that, whether it’s with personal computers or electric cars or any other innovation, there’s a chasm, a time when adoption that was going quickly among the early adopters, fails or stalls as it needs to leap over the chasm of disbelief between the true believes and the mainstream, “regular” people. And he looks at what it is that allows the technology to make that leap.
So we thought it would be very apt to frame VBC that way: what would it make to help HC leaders leap over that chams. And we did cross that chasm when I was National Coordinator. And we did cross that chasm; we went from 22 percent of ambulatory physicians using EHRs to 80 percent.
So what are some lessons that we can take from that experience and from the theory and Jeffrey Moore’s book.
What’s your read of the industry right now? Are you observing provider leaders possibly stalling out on progress on value-based contracting because of the climate of uncertainty around healthcare policy?
If you want to increase adoption, there has to be a real clarity of messaging, in particular in terms of defining the competition. When we were promoting EHRs, we were very clear that we don’t like paper; that paper kills. You can’t get more clear than that! And the advice we have to policymakers, that if they want to expand adoption of the program, there has to be clarity on there part that they want you to join an ACO. We do not like unmanaged FFS. That’s actually a message that I’ve heard from RFK Jr., Dr. Oz, and Chris Klomp, the incoming Deputy Administrator, the Director of Center for Innovation, that they do for support VBC, that they want there to be better prevention, more primary care, more chronic disease management, and that having FFS incentives leads to more fraud, waste, and abuse—and I would agree with that.
How would you interpret the tone coming from this administration? Is it one of hesitation? Because, as you argue in the article, progress must be made on the benchmarks.
I agree that there has to be continued attention to the benchmarking. Value-based care is fantastic in tis frame, saying, pay us less if we’re not succeeding—but, compared to what? It’s essential to get the benchmarks correct. In fact, one of our policy priorities is to get the ACPT correct, which set forth a third of the benchmark trend for ACOs—one-third of their benchmark was going to be set in a fixed way for five years. And it turns out that the CMS actuaries very badly misestimated inflation trends as being about half of what they actually were. So this is basically robbing a significant number of practices in the program from savings they had earned in 2024, because inflation was misestimated.
So you’re absolutely correct, if as a policymaker you want to increase adoption and performance, you can’t have people think the benchmarks are arbitrary. I’m optimistic that they will do the right thing; they have the regulatory freedom to weed out the ACTP. The Biden administration was too optimistic on inflation, and they need to fix it. So that is a valid point.
In terms of this balance between, do you try to pull or push—and Trump I—the first Trump administration—said, and I suspect Trump II will also say, if you’re not performing, you have to have skin in the game. I’m OK with that, because I think there are people who are kind of coasting in the program, not achieving much savings and not doing much work, and there’s no penalty for not performing. So creating a platform for lighting a fire under people, wouldn’t be the worst thing.
How might the adoption of artificial intelligence and other advanced analytics tools help?
We are already using generative AI as well as more traditional methods, on predictions; and the real innovation in population health is when you don’t just do everything the same for everybody, but in fact, you identify very discrete subsets of the population that will need your help, and you take advantage of that understanding. One of the things we’ve learned is that the new AI models can not only make very good predictions, but they can also explain themselves, in ways that machine learning models couldn’t. So you can ask an LLM, what is the likelihood of a patient crashing into dialysis? And it will give you a very good explanation for why this person is at high risk of kidney failure and should be reached out. To. So LLMs are improving not only prediction but explainability.
There is a somewhat technical policy issue that you and your colleagues at Aledade are deeply concerned about, and that is the ACPT, the Accountable Care Prospective Trend. Could you briefly explain the issue for our readers?
Certainly. Until 2024, the benchmarks were updated in the following manner: we waited until the year was over, and counted per-person inflation. There’s always been a circle of folks who have wanted to project an estimate from the beginning of the year, though the downside there is that the projection can be off. In 2024, CMS started to introduce ACPT for the first time.
For all ACOs in the MSSP?
For everyone who started a new contract, and that amounts to about 20 percent of the ACOs in the MSSP; over time, that number will grow.
What are we finding out?
Basically, CMS officials said, we’re going to count two-thirds of the trend and project one-third of the Trend. Two things happened: first, they ended up not giving us the Trend until November, when the year was already basically over; worse, they had to revise it in March 2025, and they still got it mostly wrong. Both at Milliman and Aledade, a number around 7-8 percent inflation sounded as though it would be accurate, but CMS made that number 3.6 and revised it up to 4.9. In effect, that put us at the minimum number, 2 percentage points off. It doesn’t sound like a lot, but we think it will cost 20 percent of MSSP ACOs around $200 million. And, given that it will take another couple of months to submit their claims, per claims submitted through February, the number was 7.2
That’s a 2.3-percent gap.
Yes, and the average savings rate for 2023, was 4 percent. So that represents one-third of MSSP savings revenue. We can envision ACPT as a concept being needed in the future. But they implemented it way too soon, and in a very crude way; not a lot of thought went into it. They took an MA number, USPCC, and said, we’ll take that number and use it raw for five years. It’s a problem.
More broadly, how do you see the MSSP evolving forward over the next few years?
I think the ACO REACH program was a CMMI innovation model—those models typically have a five-year shelf life. And what’s supposed to happen is that if elements of that model seem to be working, elements of it get included into the MSSP. So I do see the ACO REACH program sunsetting, but features like 100-percentupside/downside risk should be rolled into MSSP, or Pathways to Success. Right now, it has the basic and enhanced tracks. The basic track has 40-50 percent upside sharing, meaning if you get $100 of savings, the ACO gets $40-50 and the government keeps the rest. And very little downside; that’s the basic track. In the enhanced track, you get up to 75 percent of upside, the government keeps 25 percent. But if your costs go up, you owe the government 40 percent of that cost. What I hope will be a new track introduced into MSSP from ACO REACH, will be a full-risk track into MSSP, where you’re responsible for 100 percent of downside, but you get 100 percent upside as well.
What percentage of leaders would be interested in participating in that model?
I think half of the our Aledade ACOs might go for that, if it’s constructed well.
What would you say to those in the audience who have not yet entered in?
One of the points we made in the article is that you need simplifiers; this stuff is complicated. What I said about ACPT and tracks, it’s all complicated; and the analytics and data are complicated. That’s why we need simplifiers. We’d like to serve them with an easy button, and I need someone with the same incentives as mine. And I will trust them to my partner on this journey.
