Emergency Medicine’s Shift With Alternative Payment Models

This discussion was recorded on January 9, 2024. This transcript has been edited for clarity.

Robert D. Glatter, MD: Hi. I’m Dr Robert Glatter, medical advisor for Medscape Emergency Medicine. Joining me today to discuss a novel approach to reimbursement for emergency medical services is Dr Jesse Pines, chief of clinical innovation, US Acute Care Solutions; professor of emergency medicine at George Washington University and Drexel University; and an expert in health policy.

Also joining me is Dr Steve Schenkel, professor of emergency medicine at the University of Maryland School of Medicine, medical director of the emergency department at Mercy Medical Center in Baltimore, and also an expert in health policy.

I want to welcome both of you gentlemen to this discussion.

Jesse M. Pines, MD, MBA, MSCE: Thanks for having us.

Insights Into Alternative Payment Models

Glatter: Jesse, your article with Dr Leubitz in the recent ACEP Now really resonated because, obviously in emergency medicine, we’re trying to reduce costs where we can but also provide high-quality care. The impact of this is felt everywhere, and what we do at the beginning in the emergency department matters. The fact that Maryland has had the ability to craft such general and advanced APMs is interesting. According to your article, this dates back to the 1970s, I believe.

I’ll start off by asking you to educate our audience: What is an APM?

Pines: Basically, an APM is an alternative payment model. The operational definition is one that is not traditional fee-for-service, which tends to be transactional, where a patient and a clinician get together, there’s a bill, and there’s really no change in that bill based on the quality of care or the value delivered — that’s traditional fee-for-service.

APMs do it in a different way, and it can be done by either paying a bonus to the clinician for providing better-quality service or taking money away from that clinician for providing lower-quality service, through some sort of gain-sharing model where there may be some savings from doing things a different way. That would be shared. The highest level of APMs is global capitation, where there’s a fixed amount of money to take care of a population.

Glatter: Can you describe the mechanics of this model, including how it operates, how it works, how calculations were made, and also explain how you arrive at the moneys that emergency physician groups and hospitals can retain? Obviously, also discuss the behaviors that they can engage in and practice to make this APM work.

Pines: There’s variation in physicians’ decisions to admit patients to the hospital. Work that I had done years ago demonstrated that there was two- to threefold variation within and across hospitals in terms of this decision that emergency physicians make. That was really the beginning of the basis for this: Can we start leveraging some of that variation, get clinicians to put in evidence-based protocols, and then provide some shared savings back to the clinician for appropriately reducing admissions where the patient may not need to be in the hospital?

Basically, the architecture of this is that it focuses on Medicare fee-for-service patients only because that’s the group that is regulated by the group at the Health Services Cost Review Commission (HSCRC) of Maryland. It hones down to diagnoses where there’s variation in the decision to admit. That includes diagnoses like chest pain, pneumonia, and syncope, and takes out things like severe trauma, heart attacks, and strokes. It also takes out things like ankle sprains and abrasions where there’s just not going to be much variation.

It basically takes the total cost of those visits for a baseline year, and in the case of this 2019 program, that cost goes out to 14 days. Then that’s compared to the cost in 2023 after inflation adjustment for those same diagnoses. The concept here is that, if you’re able to appropriately lower hospital admissions for those conditions where there’s evidence to do so, the physician group would get a share of the overall savings based on the total cost of care. That savings is 20%, 50%, or 80% depending upon how efficient the baseline period was for that group.

It’s a complicated model, but it’s around lowering admission rates where possible, comparing a baseline to a performance period, and ultimately rewarding high-value care.

Glatter: There is something to be said about clinical gestalt. Certainly, we can use pathways for dealing with chest pain, the HEART score, atrial fibrillation pathways, and so forth, as guidelines. In a bell-shaped curve, maybe these critical pathways can help in 80%-85% of cases. Altering physician behavior is really the crux of it, and I think that speaks to where these APMs are valuable.

Steve, would you agree with that?

Stephen M. Schenkel, MD, MPP: I think that’s true, and I’d expand on it, Rob. When we talk about an APM, it can be of many different things. An APM could be as simple as a negotiation for a fixed fee per visit from a given payer.

In this case, what we’re talking about is something much broader in terms of global payments for specific conditions and the care associated with those conditions. Part of this is trying less to alter physician behavior than to narrow physician behavior, so that we’re practicing in similar mechanisms, combined with working with the structures around us to provide the backbone and support that we need to make other decisions.

For example, you mentioned that Maryland is different, and Maryland is different. We’ve worked under an APM for hospitals for many decades now. One of the challenges we’ve run into as that model has advanced is that the incentives or goals for individual clinician groups may not match the hospital incentives or goals to the degree that an alternative payment mechanism can align those goals. Not only are emergency physicians given reasons to think of alternative mechanisms of care, so are our hospitals and so are our community resources. Ideally, the alternative payment mechanism will align us across hospitals, across communities and across physicians.

Comparing APMs With Kaiser’s Integrated Care

Glatter: How does this type of model that you started in January 2023 compare with Kaiser’s model?

Pines: There are three different basic approaches. One is to give a bonus or take away based on an individual quality visit, like, “Did you do X process?” Another is shared savings, which is the program that’s in place. The third is global capitation, which I think is really the final destination for this, where, like in the Kaiser model, they have a fixed amount of money to deliver acute care services to a population.

To do that as efficiently as possible involves not just seeing patients in the emergency department but doing other things around that that can direct patients to the right location at the beginning, making sure things happen efficiently in the emergency department, and making sure that people have good transitions of care after the emergency department.

Really, the next iteration of this that we’re trying to put together is a global budget model for emergency departments that would do just that. It would really align the incentives with the emergency group and the hospital toward this broader population health goal.

Glatter: One issue is having the services set up, for example, if we’re not going to admit a patient. That’s what Kaiser can effectively do, and has done, to reduce their expenditures. They can pivot quickly away from admissions, do outpatient care, and set up telehealth services. The question is, why can’t the bulk of large healthcare systems embrace this type of care that Kaiser has shown to be very beneficial?

Schenkel: What are the goals? What is the mission? What is the funding? Not to be too simplistic about it, but it’s that alignment of incentives to say that this alternative destination is important enough to develop it. None of this happens quickly, right? There’s a change in pattern that then is reflected by successful disposition and follow-up. I want to know, if I discharge someone, that they will get followed up and there’s an easy way to get there.

In setting up programs like this, there’s a tremendous emphasis on that. I want something I can access 24 hours, 7 days a week to make sure it’ll happen tomorrow. I need to be able to do it in a mechanism that, when it’s a 3-day weekend, I have some alternative from Friday to Tuesday — the emergency medicine mindset of, it is important to have things in place regardless of the initial time and day. It’s not so straightforward to do, but the answer is “yes” if it’s important enough.

Pines: One of the things that really differentiates any program that we would launch in Maryland and Kaiser is the degree of alignment. Steve got at this. Wouldn’t it be great if you had three entities (your insurance division, your physicians, and your hospitals), all of whom had equal incentive to move toward delivering the best product available to the patient? That really is what an integrated healthcare model does, because you have a fixed amount of money, you’ve got the various players in there, and ultimately everyone benefits if things happen more efficiently.

In traditional fee-for-service, you’ve got different parties broken up in pieces, which really creates much of the fragmentation in services and the misalignment. If the patient does need a rapid follow-up appointment with a specialist, say an orthopedist, that orthopedist is not explicitly incentivized or held accountable to actually have that appointment happen unless you’re in an integrated system.

We need to get closer to full integration. Global budgets in Maryland for hospitals is only one piece. Adding global budgets for clinicians on top of that gets us a little closer, but still, it’s not that full integration because that’s only the clinicians and the hospital itself, and it doesn’t apply necessarily to the outpatient environment, where they may not be held accountable for patients to be seen in a timely way. All these models are inching closer to that global capitation, but you’re absolutely right — it doesn’t get us 100% there.

More on Maryland’s APM Program

Glatter: This is not in any way the carrot or the stick, but physician behavior in admitting, in terms of high admitters, is one aspect to the success of an APM model, and ability to discharge patients is part of this whole exercise. We need to all work together. That’s really the message. We need to be unified in terms of reducing expenditures.

Steve, in terms of physicians that are engaged in this APM at Maryland, do you have to elect in? How is it that ER physicians enter into this? Is that electively or is it the entire emergency department?

Schenkel: Any of these types of operations is going to be done by a group. It may be a single hospital group, depending on how it’s created, or larger groups. You point a finger at something very important, which is that so much of what is essential in this comes into the details — for example, whether billings are being done exclusively by the group or by the hospital or by a larger consolidated group of emergency physicians. Because these programs tend to be group focused, they’re going to be a whole group joining and not individual clinicians. I can’t speak across Maryland, but I think relatively few of us are independently billing as individuals, so I’m not sure that the individual approach is workable.

Glatter: According to the article, 600 emergency department physicians are part of this APM that was launched on January 2023. How was it that these 600 got selected? Was this an opt-in/opt-out option that was presented to emergency department directors throughout Maryland or healthcare systems?

Pines: The program was brought together by stakeholders which included HSCRC, the Maryland chapter of the American College of Emergency Physicians (ACEP), MedChi [Maryland’s state medical society], and Maryland Hospital Association. Once the program was created, it was then offered as a voluntary program to any physician in the state who wanted to join the program. It is a physician-level program; however, we work in groups within a site. The physicians have to sign what’s called a care partner agreement in order to join the program, and 600 physicians did for the 2023 program. What is being measured is the site-level admission for those included clinicians.

Again, it’s a voluntary program. It’s also what’s called an advanced alternative payment model (AAPM) where, for the Medicare program, you either can be in a merit-based incentive payment system (MIPS) or you can be in an AAPM. In an AAPM, you don’t have to do a lot of the MIPS reporting and there’s no potential downside.

The other benefit of the AAPM — in 2023, at least — is that there’s a 3.5% bonus for all Medicare fee-for-service patients, for physicians who are participating in the program. Regardless of whatever that potential upside is, with the shared savings, you get 3.5%. There’s no potential downside risk of MIPS.

In 2024, we’re still trying to decide on what the AAPM bonus is going to be. I think the estimate is likely going to be 1.75% of Medicare with no downside for MIPS, and I think that will likely be codified going forward after that. I think that’s still outstanding with the congressional legislation.

Maximizing Benefits: Tailoring APMs to Hospital Volumes and Various Settings

Glatter: In the high-volume hospital systems, the physicians enrolled in these AAPMs in this program, for example, would certainly benefit more just on volume alone. The smaller hospitals with lower volumes wouldn’t see the economic benefits, I guess, personally to the physicians.

Pines: There would be less of a bonus because the bonus itself is correlated with the number of Medicare patients you’re seeing, the total cost of those Medicare patients, and whatever potential reduction you see. When it comes to a number like the 14-day total cost of care, that ends up being a very large number. Even a relatively low percentage reduction in that number can be a pretty substantial number, even for a small emergency department.

I also want to mention that there is what’s called a 3% hurdle rate. Once you lower costs by 3% (ie, inflation-adjusted costs), then you basically get shared savings on the upside. Let’s say there’s a 5% shared savings on total cost of care for 5000 Medicare patients at a smaller emergency department; that can still be a substantial number.

Glatter: The 14-day aspects of care that are part of this program interest me because the downstream care provided after our encounter is really not necessarily all in our control. That unification and teamwork is important. How is that all mapped out?

Pines: The architecture of what’s called the Episode Quality Improvement Program (EQIP) in Maryland requires use of a total cost of care as the outcome. When it comes to emergency patients, where there’s variation in the decision to admit, the big variable there is going to be whether or not you decide to hospitalize the patient, regardless of what happens in that tail period. What we found is that you’ve got someone with chest pain and you admit them, and they’re in the hospital for 2, 3, or 4 days vs being lower risk, and being discharged and being seen in the outpatient system. That outpatient pathway is going to be much less costly.

However, you’re absolutely right. For someone that we discharged from the emergency department, being responsible for up to 14 days of total cost of care — that was some of the pushback from emergency medicine. How can I be responsible for that? Mathematically, it works out because the way that we can control that is through something that we can implement an evidence-based protocol for.

Glatter: Steve, to employ such a model at your institution, would that be feasible in terms of, say, downtown Baltimore and the indigent population, patients without resources or care, with substance abuse and trauma? Social determinants of health play a role in this model working, clearly.

Schenkel: I’m not sure there’s a question of feasibility as necessity. I say that from two perspectives. I return briefly to a question you asked of Jesse, which is the decision to join the EQIP model. No decision is made in isolation. It’s not a decision of, do we do an APM or not? The question is, do we do this? Do we do nothing? Do we do MIPS? That made for a fairly simple decision, especially given that the APM aligns with hospital goals.

That gets into the second question, which is, how do you do this in a downtown setting? How do you do this in any setting? The answer is that we have to be pointing in the same direction. I’m not going to be able to directly influence, for example, every cardiologist in Baltimore City.

However — and this is especially true in Maryland — where we live, in an APM for the hospitals, they’re all hearing a similar commentary about trying to keep patients out of the hospitals, trying to take care of them at home, and trying to direct them toward the most appropriate care in a timely fashion. If I have those direct connections, great. Even if we don’t, presumably we are starting to move in the same direction. Obviously, all of this will evolve and is continually evolving, but we have to have that coordinated effort. That is true whether or not we’re part of the model, simply because of the larger umbrella of reimbursement that we’re working under.

Charting the Future: Sustainability and Expansion of APMs

Glatter: Jesse, you had mentioned a 1.75% change in 2024. Is the sustainability of this model something that’s already in place or is it up for discussion? Where do you feel this is going from a federal or state level?

Pines: I think it depends upon how successful this model is in the first year at lowering costs, and I think that a number of things are going to hinge on that. One is if the groups are able to actually lower costs. Two is participation. I don’t have the numbers for how many physicians are in the program in 2024. I do know that many more groups decided to join compared with 2023. I think that is one marker in clinicians deciding to choose this program.

The architecture of the program is based on shared savings, which is, by definition, a time-limited program and is really the problem with any shared savings model, because you have a baseline and a performance period. You can improve on that and then you can reset the baseline to a new performance period. You can only do that so many times before you’ve squeezed out all the efficiency. That’s been seen time and time again in accountable care organizations, other APMs.

I think the final destination for this will be a global budget. Will this serve as a model that could be used by the federal government, by the Centers for Medicare & Medicaid Services (CMS), or in other parts of the country? It has many similarities to the ACEP acute unscheduled care model. If this is successful, does CMS pick this up? I would say there’s a reasonable likelihood of it being offered more nationally. Again, it’s time limited, capitation, global budgets — that’s really the only long-term model that’s going to get full alignment.

Optimizing Emergency Admissions: Strategies and Considerations

Glatter: In terms of physician behavior in admitting, have you seen pressure in any way, like physicians that admit more patients for certain diagnoses in this 500-plus group of diagnoses?

Pines: That’s exactly how we have implemented these programs, and we also have a history of having done this for a few years prior to 2023 in other markets within US Acute Care Solutions very successfully. Our model is one of identifying people who are outliers, and there are tremendous outliers on both the high end and the low end when you look at this. Two is making sure that the data are right. Three is sunshining the data and showing people that they’re an outlier. That has a very large effect. I like to say that 95% of physicians think they’re above average. Telling a physician that they are not above average often does not compute. People may think more carefully about how they’re practicing compared with their peers.

The final area is having those evidence-based protocols be accessible. Many physicians know about the protocols; whether they are using them in the best way possible is really the question.

Glatter: Admissions sometimes is necessary for certain types of patients that don’t have the resources or the ability to follow up. We have to account for that in these models. At larger urban centers with a more indigent population, that certainly is a concern.

Schenkel: Regarding the concept around admission, you’re absolutely right. Some people need to be in the hospital. There is a question, though, of who truly benefits from hospitalization and how do we optimize that benefit? There’s an underlying question in there of what is the purpose of the hospital? That’s philosophical and a rather large question, but we do need to recognize it as a question, especially as we are looking at things like hospital-at-home and how to operationalize quite extensive care for people in their homes.

The other is a perspective on reducing hospitalization and physician variation. There are two ways to look at what we do when we make efforts. I want to highlight what Jesse said, that we tend to focus on the person who admits many people, but there’s nothing to say that the person who’s at the other end and admits the fewest is right. Some people need to be in the hospital, we think, and maybe some aren’t going in.

One approach is to say, I want to shift the whole curve so everybody admits less. The other approach is to say, I want to get some logic and evidence behind this so I’m narrowing that curve. Maybe after I’ve narrowed that and I have some better understanding, I’ll shift it. Often, though, that first step is just narrowing and trying to see if we’re actually practicing in similar ways. I think we put a large amount of effort in that. That’s much less threatening as well when we talk about practice and practice variation.

Finally, there’s the economics of all of this. One of the attractions of fee-for-service is that it’s very simple. You do something, you get paid for it. It’s like going to a store. That transactional concept is very simple. The moment we move into other budgetary approaches and models, they become far more complex.

One of the underlying streams in our conversation is that the details matter greatly. Jesse’s example of the time-limited comparison is an incredibly important one and one that can get lost easily. The details really get honed as we move along, and that change in honing is very important as we understand better what we’re doing it and how to improve what we do.

Pines: My final point here is that the involvement of emergency physicians in designing these models is so incredibly important. When these models that apply to the emergency department are designed by economists or folks outside the emergency department, there’s often misunderstanding about what we do and why we do it.

The early success in some of these programs in Maryland is that this was designed by emergency medicine for emergency medicine. I think that really has to be the model going forward for success. We need to be intimately involved in how these programs are designed and run because there is a number of potential unintended consequences that have already occurred in the state of Maryland.

For example, they did not consider how global budgets for hospitals have impacted flow in the emergency department and prioritization of making sure that people are seen in a timely way. We need good physician leadership, understanding of the big picture, and stakeholder alignment to make any of this work.

Glatter: The more we can do to take better care of our patients, but also work on the economic side and try to ease burden on our systems and our hospitals, and make it better for everyone, I think that’s our common goal. I want to thank you both for an informative discussion on APMs. You’ve been very helpful.

Robert D. Glatter, MD, is an assistant professor of emergency medicine at Zucker School of Medicine at Hofstra/Northwell in Hempstead, New York. He is a medical advisor for Medscape and hosts the Hot Topics in EM series.

Jesse M. Pines, MD, MBA, MSCE , is a clinical professor of emergency medicine at George Washington University in Washington, DC, and a professor in the department of emergency medicine at Drexel University College of Medicine in Philadelphia, Pennsylvania. .Pines is also the chief of clinical innovation at US Acute Care Solutions in Canton, Ohio.

Stephen M. Schenkel, MD, MPP , is a professor of emergency medicine at the University of Maryland School of Medicine, and chief of emergency medicine at Mercy Medical Center in Baltimore. Schenkel was previously the president of Maryland ACEP. He currently co-directs programs in administrative education and lecture tracks in behavioral health and morbidity and mortality. He is also an associate editor for Annals of Emergency Medicine and JAMA.


Leave a Reply

Your email address will not be published. Required fields are marked *