What does the journey into value-based healthcare delivery and payment look like, at the patient care organization level? On Wednesday, two senior executives from the MetroHealth System, an integrated health system in Cleveland, shared some insights into their organization’s journey so far, at the Health IT Summit in Cleveland, sponsored by Healthcare Informatics.
In their keynote presentation, entitled “MetroHealth’s Road to Success: The IT Approach for One of 30 National Successful MSSPs,” Nabil Chehade, M.D., senior vice president of population health, and David Kaelber, M.D., Ph.D., CMIO and vice president of health informatics at the 731-bed system, which is anchored by MetroHealth Medical Center, shared with their audience both the high-level strategic components of their accountable care work, and the strategic and practical IT elements involved. MetroHealth was an early participant in the Medicare Shared Shavings Program (MSSP) for accountable care organizations (ACOs), and its leaders already have a host of learnings to share with colleagues across U.S. healthcare, regarding the challenges and opportunities in this area.
Starting at a U.S. healthcare system-wide level, Dr. Chehade began the joint presentation by contexting MetroHealth’s work in this area. Speaking of the overall U.S. healthcare system, Chehade noted that “We spend 17-18 percent in this country, compared to an average of 10 percent” among other advanced industrialized nations. “Do we get more out of spending more? Actually, we don’t. And there is county-level data that shows you how every county in this country is doing” in terms of clinical and health outcomes. “I can tell you that Ohio doesn’t do very well. More importantly, with regard to this county, Cuyahoga County, which is home to some of the best patient care organizations, on a national level, still, this county’s health outcomes are some of the worst.”
In fact, Chehade said, “What we have built throughout the years is a system that doesn’t work. How can we spend all this money, without a lot to show for it? If you were running a small business yourself, you would never run it that way? But somehow, we allow it to happen in our system.” Thus, the burning platform for population health. “What is population health? Do the right thing for the patient, give them better care, good service; but at the same time, there is only a limited amount of dollars, so how can you use those dollars to deliver the best kind of care? Population health medicine is a different way to deliver care,” he said.
And, Chehade noted, the Affordable Care Act, passed in 2010, helped to push providers forward towards value-based healthcare, with its numerous provisions for accountable care and bundled payments. In that, he and his colleagues at MetroHealth have been eager to move forward with alacrity into value-based care. That said, he noted that “It’s a fine balance for any organization to live in both worlds, the fee-for-service world and the value-based world. Is there a way to do it? Of course there is. Where there’s a will, there’s a way.”
Further, Chehade said, “Why do we believe we can do this? The first pillar” in his organization’s efforts has necessarily been “our patients and our providers. We still serve a lot of less-fortunate population. Close to 40 percent of our patients are Medicaid patients. And just because we had so many Medicaid—and uninsured, before Medicaid expansion in Ohio—patients, we had to work hard to successfully serve those populations. But that has led us to be able to recruit clinicians and others who share that sense of mission.” Meanwhile, he noted, “The march from fee-for-service to value-based will never happen if an organization’s leadership is not fully committed. Our leadership committed to branding MetroHealth’s transformation as a transformation to population health. That’s what we preach and live every day, and that’s very important.”
In that, he said, “We started by building up our patient-centered medical home. Then came the pre-expansion waiver under Medicaid. MetroHealth was a demonstration site, and we did very well. Then, in 2014, we entered MSSP Track 1. In 2015, we entered with our largest HMO payer for Medicaid, with 70,000 lives. And then in 2016, when I arrived at Metro, we established the Population Health Institute. And soon we decided to take not only upside risk, but upside-downside.” For example, he noted, “In the Medicaid CPC Ohio program, we’ve taken both upside and downside risk for 100,000 lives.”
As a result of those initiatives, Chehade said, “Looking at our core population, 75 percent of patients are in a value-based contract; and in Medicaid, it’s nearly all patients in some kind of risk contract. How do we do this? We start with payment reform. We believe we’re the only organization in the country where any contracting in value-based payment is first run through our department [population health], and not the CFO; and we believe that’s very important, because it aligns the flow of money. Then we start segmenting the population, creating patient registries. We look at our community of patients and how we’ll engage them. And how can we transform ourselves. And in every single step of the process, we need to engage our IT and informatics.”
Moving into the practical, Chehade told the audience that “Our Population Health Institute has dedicated clinical leadership councils. One council is responsible for quality outcomes, specifically as they relate to our value-based programs.” In that, he reported, “We have a dedicated physician and dedicated nurse, leading a multidisciplinary effort, engaging our clinics, hospitals. Then we have a cross-continuum care optimization team; that team goes after the efficiency and the dollars; looks at readmissions, looks at how to decrease total cost. And finally, we have a market solution team, which aligns the flow of money into the organization, to be able to navigate that delicate balance between the fee-for-service payment and the value-based payment. It’s a continuum.”
Also, he noted, “We have a population outcomes and evaluation team; we bring to them challenges into the claims data, challenges into the risk stratification model; and we don’t go to a vendor for this, we create those things internally, and approach our patients in a different way.”
Leveraging data and IT for population health success
Dr. Kaelber then spoke, connecting the MetroHealth’s broad population health management strategy to the data and IT components of the organization’s success to date. “We’re leveraging the electronic health record, which makes this possible,” Kaelber said. “We’re a big Epic shop, so we’re using the tools in the Epic sandbox. And for this activity, we’re limiting ourselves to the tools in that sandbox, but there are a lot of them.” And, he noted, “We’re putting those tools together in specific ways geared towards population health and risk contracts. What I’ve learned in the past five or six years is that you have to start by identifying your patients at greatest risk. And that starts with a registry,” he said.
“Sometimes you know a priori who your highest-risk patients are; but as a Medicare ACO participant, you’re given retrospective attribution, while you’re trying to do prospective work,” Kaelber noted. “So we’ve been able to determine by proxy who our highest-risk members are. And It’s great six months after the fact to know how we’ve done on some HEDIS [Healthcare Effectiveness Data and Information Set] measures, but that’s not particularly helpful as we work in risk. So for real-time clinical practice, sometimes, we use proxies.”
One area of particular complexity, Kaelber noted, is around the CMS [Centers for Medicare and Medicaid Services] hierarchical condition categories, or HCCs, with the HCC model generating a risk score for each beneficiary that summarizes each beneficiary’s expected cost of care relative to other beneficiaries. In order to be successful in shared-savings programs, he said, “It’s important to know the expected cost per patient, compared to the actual amount spent. And how does CMS know what to expect? It’s based on the diagnostic codes we give them; so we have to accurately give them those codes. And just the way the program works with CMS is, somehow, CMS assumes that on January 1 of the year, patients are totally healthy. And so even if you saw a patient on December 31 with metastatic prostate cancer, renal failure, congestive heart failure, diabetes, and everything else, maybe they saw an amazing doctor on that day. Of course, that’s not real, but you have to make sure to tell CMS that you’re still managing those current disease states. And so we can look at past medical history, encounter diagnoses, even BMI. And that’s not a diagnostic code, it’s height and weight in the EMR. And you have to identify gaps—you have to make sure that you put in all these HCC modifiers, basically diagnostic codes, so that CMS understands how sick these patients are.”
The key fact in that, Kaelber said, is that “Our electronic health record has mined all the current diagnoses, and also analyzes conditions that might seem to be missing.” What’s more, he said, “Here’s another example: maybe somebody had a diabetic foot ulcer last year; and it is possible they no longer have that. But you need to enter the diagnostic code. And beyond that, you also have to document that not only did you bill for a diagnosis, but you did something during a visit, around that diagnosis.”
In terms of identifying patients who will need interventions, Kaelber said that, “Per clinical decision support, we just want to make doing the right thing as easy as possible for physicians and everyone on the care team. And in that,” he said, “patient registries are key. We’re making sure that CMS spends appropriately based on how sick our patients are. The other element is all the preventive health measures. So what we’ve done is, in our EHR, we’ve basically built in all the 32 quality measures amenable to point-of-care interventions.” Other important elements: “Getting patients in for their annual wellness visit, and getting them immunized, etc.; any time a patient comes in, the system shows the physician what’s going on with that patient, because that’s a key moment for intervention.”
Another element in this is patient engagement, Kaelber noted. “Well ask patients, why did you come in today? And sometimes, they’ll actually say, well the personal health record made me do it. And that shows that our outreach to patients has been effective,” he noted. Indeed, he reported, “As of last month, 50 percent of our patients—150,000—have an active MyChart account. One in three, even in the MSSP, has access to this tool that we’ve leveraged.”
Meanwhile, on the operational side, Kaelber said, “Another piece that’s been key to our success has been the use of dashboards. It’s great to do things around registries and diagnostic codes for billing and measures for quality of care, but we also need to be measuring what we’re doing. So about a year ago, we put in a bunch of dashboards—both provider-level dashboards and executive-level dashboards. And we’ve actually allowed this for everyone, to begin at the system level, and making them able to drill down to the individual provider level.”
Indeed, he said, the level of detail of drill-down involved, including with quick-reference coloring of data in red, yellow, and green (to indicate the level of performance of a data point), has been particularly helpful. “And these measures are recalculated either daily or monthly, depending on the measure,” he noted. “And knowledge is power: if we don’t see a particular individual provider moving in the right direction, we can intervene before it’s too late.”
Of course, MetroHealth leaders inevitably have had to guide their physicians into acceptance of transparency around physician performance measures. “What if Dr. X and can go into the system dashboard and see how Dr. Y is doing? Our view is, that’s good. That’s transparency; and data is power. There was a little bit of internal discussion” when data on individual physician performance was made visible to all physicians and to others in the organization, “but at the end of the day, we said, we’re going to allow anybody to look at anybody else’s data. And you can drill down by provider, location, facility; in a couple of clicks, you can slice and dice the data however you’d like. And individual providers can drill down on individual patient lists. If I’m not at the right level on a certain measure, I can look at which diabetic patients I’ve not seen whom I should see, for example. With real-time, easy access to all the data, better care can be provided,” he emphasized.
Into predictive analytics
Kaelber reported that “We’ve begun to leverage predictive analytics in this space. We’re looking at things like readmission rates,” he noted. “One approach is to say, wow, everybody who is admitted has a risk of being readmitted within 30 days; but actually, not all those patients are created equal. And if we spend all of our time and energy focusing on everyone, that’s probably not going to be helpful. We need to focus on the higher-risk patients. So we have a tool within Epic that gives us a readmissions risk tool. I can look at every patient in my care who is at risk for readmission. And you can look at every factor involved. Some of those factors might be modifiable, some not, inside that readmission risk score. And we’re putting extra resources into the population that’s at extra risk, based on our readmissions risk score.”
One thing builds on another, Chehade emphasized. “We’ve been building managerial/care coordinator dashboards, to look at which patients are at high risk, and who they are, and we’ve been working to subdivide them by risk. In fact,” he said, we can identify individual patients at higher risk, so that our care coordinators, case managers, social workers, can be deployed effectively, and can approach patients who are high-risk and rising-risk.”
That kind of work supports the overall effort involved in making accountable care a successful proposition at MetroHealth. With regard to the MSSP program, Chehade said, “We have about 11,000 patients in that population. We have upside-downside risk for a total book of business that’s worth $110 million, with a potential downside of $18 million, and a potential upside of $20 million. So we risk-stratify every single patient. And we have a high-risk group and a care coordination program for them, and a rising-risk group and a care coordination program for them. And there are low-risk patients, and patients for whom we don’t have much data, and we have navigators for them. But nothing is left to chance.”
Meanwhile, he noted, “Our sickest patients are responsible for 30 percent of the cost; and we have a specific program for them that’s a ‘red carpet’ program; and even within that group, there is a highest-risk group, with 300 patients, and we employ different levels of intensity, which we call octane levels. What is a high-octane program? It’s patient-centered medical home on steroids, a multidisciplinary team. If you’re on high octane, you’re most likely being touched at least twice a week by the care team, being visited in your home. You probably have social services, a care coordinator, a case manager,” and so on.
“Can you prove that this works? Our red-carpet program began a few years ago, but really matured last year,” Chehade said, pointing to a chart showing a nearly 80-percent reduction in ED visits among the high-risk patients being care-managed, and a 12-percent reduction in all-cause inpatient admissions among high-risk patients. “When we started with our MSSP in 2014, we actually overspent by about 2.1 percent. The following year in 2015, we did pretty good, 1.7 percent, but not good enough to get a penny of shared savings. And in 2017, we committed to upside-downside risk. We believe we’ll achieve more savings than in the prior year,” he said. In any case, he emphasized, “We cannot afford to continue the way things are going in this country. The driver has been CMS and Congress, because they understand the cost issues. Things are changing in Washington.” And the cost trajectory and other factors are accelerating change now, he concluded. “We’re living a healthcare revolution.”
