Fowler Describes CMMI’s Challenges, Work on New Payment Models

Dec. 14, 2022
Elizabeth Fowler, director of the Center for Medicare and Medicaid Innovation, engages in a lively virtual Q&A with the Penn Medicine community in Philadelphia

On Dec. 12, Elizabeth Fowler, Ph.D., J.D., the director of the Center for Medicare and Medicaid Innovation (CMMI), engaged in a lively Q&A with the Penn Medicine community in Philadelphia moderated by Rachel Werner, M.D., PhD., the executive director of Penn’s Leonard Davis Institute of Health Economics. Fowler began by talking about several projects currently on CMMI’s plate.

She noted that President Biden recently signed into law the Inflation Reduction Act, which aims to address prescription drug prices. CMS is busy implementing that law. In addition, the President signed in October an executive order directing the Innovation Center to outline possible new models that have the potential for lowering drug costs and promoting access to innovative drug therapies for beneficiaries in Medicare and Medicaid. “We are busily working on that,” Fowler said. “The report is due essentially the third week in January. We're looking at targeted model concepts to increase affordability of drugs and increase access to those novel therapies, so stay tuned. We're thinking about prescription drugs as well in a very substantive way.”

What follows is a summary of the exchange between Fowler and the Penn Medicine community about work on alternative payment models at CMMI.

Werner kicked things off by asking how the experience of working through the pandemic has shaped how CMMI is thinking about value-based purchasing and what it means for the direction CMMI is going.

Fowler: I think that what we saw during the pandemic was that providers who were in value-based care models or in value-based care arrangements were more resilient. They had a more stable and predictable stream of income. They had created those links with community-based providers; they knew more about their patients, and where those patients were in their healthcare journey compared to providers who were solely dependent on fee for service. I think what we thought was that coming out of the pandemic, there might be a rush to sign up for new models. But I think almost the opposite has happened. I don't want to be too pessimistic, but I think we're still in the pandemic and a lot of the providers who have really stepped up and delivered for the U.S. population — thanks to all of them who were part of that response — but they're not in a position now to take on additional risk and jump into models. We're still seeing a lot of challenges, particularly in rural areas, particularly with workforce. That's made a lot of providers very reluctant to go down this path. If anything, it has made them a little more risk-averse. So if they were nervous about jumping into taking downside risk, the pandemic has certainly not encouraged them to do so. We recognize that, and I think when we're outlining our strategies, we are trying to take into account that at some point we will be over this hump and back in a position where these models are more viable. We're preparing for that future.

Werner followed up by asking about whether CMMI and CMS are considering deploying more mandatory alternative payment models.

Fowler: There's legitimate disagreement. I happen to think that taking downside risk and mandatory models is a way to help accelerate our movement toward value. Some disagree with that and don't believe that we need mandatory models — that eventually we'll get there on our own. It's an active debate, not just publicly but also even within the administration. I think we need to get there.

A follow-up question touched on the radiation oncology alternative payment model, which was supposed to be mandatory, but has been delayed a number of times.

Fowler: It was blocked by Congress, which stepped in and delayed it. We made some adjustments to it in response to some of the industry stakeholder asks and their policy input, and then we put out the model again. It was going to be a mandatory model, so we had to go through rule-making. Congress again stepped in at the end of December last year and delayed it again. I think the challenge for us is we have to at least demonstrate that the model could break even, if not save money. The changes that the industry is asking for would make it impossible for us to get through the actuaries’ screen. We don't want to put something out that's just going to get blocked by Congress. In response to the last legislation that passed in December last year, we put out a rule that delayed the model indefinitely. If Congress wants us to take it up again, and they're willing to let us test a model that wouldn't necessarily save money, but it might improve care, they need to instruct us and tell us to do that in statute. Otherwise, we just can't make the numbers work.

An attendee asked about the connection between payment models that CMMI puts forth and the growth of Medicare Advantage. Is the reach of models diminishing over time as more and more Medicare enrollees are enrolled in MA and is that a concern in terms of the levers that CMMI has to push on value-based purchasing?

Fowler: We need to think about how all of these payment rates are set and how we set the baselines. If the fee-for-service program is shrinking, but that's the program we're using to set these benchmarks for performance, we need to think about the implications of that. We've been giving some thought to how to set benchmarks, how to determine whether those savings have been realized and how to reward those savings.

Another attendee asked why so few alternative payment models proposed by physician groups to the Physician-Focused Payment Model Technical Advisory Committee (PTAC) have actually been tested.

Fowler: When it was announced that I had this job, I got together with a PTAC member, and they gave me an earful for about an hour about how frustrating it was that the Innovation Center wasn't taking these ideas. They would do all this work and get all these ideas in and then they would just go to die at CMMI, and it was very frustrating. When I got here, the team said that you do see aspects of those ideas and proposals incorporated into models but not wholeheartedly and not whole-cloth. Given what I've explained about our clearance process — it takes 18 months to two years, the chances that somebody comes with a model idea that can be taken and implemented is probably slim to none. What we have done, though, is tried to work really closely with PTAC since I've been here on a shared goal for what we want to make sure that we're testing. We are using their platform to reach out to physician groups to get input on very specific ideas, like our specialty strategy, and thinking about some of these stickier questions such as how to incorporate equity into models. They've been great partners in thinking through some of these tough questions that we’re mutually trying to tackle.

A clinician from the Penn School of Nursing who studies models of nursing care delivery in primary care and other community-based settings said that she has found that the financial benchmarks and multi-payer alignment are just one side of the coin in contributing to the success of ACOs. Innovations on the delivery side are just as important. She gave examples of team-based care, care coordination and behavioral health and substance use alignment in primary care. She asked about CMMI priorities in advancing some of those delivery goals.

Fowler: You've described features that we want to see all beneficiaries have access to. We're trying to make sure that our models going forward capture those aspects of what we consider to be high-performing healthcare, with an important role for making sure there are nurse coordinators, and we're using providers to the top of their license and nurse practitioners. Those are all very important aspects of access and availability of services and making sure that we're doing what we can particularly for frail, elderly beneficiaries to stay at home and be able to stay at home rather than going into an institution. You're describing aspects of what it is we're essentially trying to achieve.

An attendee asked what sorts of models are deployed right now in the area of mental health and whether Fowler could speak to CMMI’s plans related to HHS’ long-term goals around mental healthcare and mental health integration.

Fowler: Behavioral health is a strong, huge priority for the administration and certainly for the CMS Innovation Center. We are working on a stand-alone behavioral health model that will look at tactics and approaches and strategies for integrating mental health and behavioral health with other aspects of primary care. We're also looking at integration as a component of a primary care model. Thinking about that integration is really where we're putting a lot of effort right now, but we also need to make sure that we're not interfering with other efforts that are aimed at improving mental health access across different populations. We need to make sure that that we're not running afoul of other efforts ongoing in the agency and the department.

An emergency department physician at Penn asked whether CMMI was discussing extending hospital-at-home programs or creating a demonstration program.

Fowler: I think we wouldn't create a specific program around hospital at home, but we could incorporate the waivers as part of a model. And, you know, in the specialty care strategy, we can take a look at that as part of a waiver. We've had several meetings within CMS on who owns hospital at home or acute care at home. We're not the owners of that, but we do have a role to play and certainly can help facilitate some of that work. We've also been providing technical input to Congress if Congress decides they want to expand and extend that waiver capability. I will say though, I'm not sure that paying 100 percent of what you would otherwise be paid if those patients were inpatients is the right payment amount. There's certainly still a high cost, but it's probably not the same as in an institution. But we don't know what that dollar amount should be. Is it 80 percent or 70 percent? So I think our preference, if it's going to continue through an act of Congress, is to have a limited time until we're able to study the details and specifications we might recommend for a permanent basis.

A critical care physician who runs a research center at Penn focused on serious illness care delivery asked if Fowler could share any thoughts related to outcomes in serious illness care that they use to train models or the outcomes that they are wishing to promote in the context of value and particularly whether there is opportunity to collaborate with academics in thinking about novel outcomes that you might use in this space. He said it is important to measure and promote the right things and it's not always so easy to do in serious illness care.

Fowler: I would love to talk further about what measures we should be looking at. We tried to implement a seriously ill population subset of the Primary Care First model and ended up not being able to get it over the finish line. We have the high-needs population track of the ACO REACH program, and there we do look at readmissions and some of the standard measures that you might consider, like keeping patients at home, not going in for unnecessary care, but we'd love to explore potential other models. I've done a lot of site visits over the last few months. And one of those was with a high-needs ACO, which was also giving us recommendations for the use of hospice care and whether they were getting adequate and appropriate hospice care, but not overusing hospice care.

The last question was, given the lag between coming up with concepts for innovation and the 18-month to two-year cycle to actually implement it, does CMMI find that the problems they are addressing have changed or shifted and their solution doesn't really fit anymore? How often is that a problem?

Fowler: I think it's a very real problem, and as value-based care continues to grow, it's harder and harder to come up with a discrete model test where you can say “Yes, this made a difference in reducing costs or improving quality compared to the counterfactual.” We have undertaken an initiative, our evaluation team and our learning and diffusion team, to think about measuring success differently and thinking about framing it in terms of transformation. Are we actually making a difference in transforming the system and really using a different approach to determining whether or not we've had an impact and looking at different metrics? So we'll see how far we get with that, but you're putting your finger on a really important issue.