Jonathan Weiner, Dr.P.H. is a professor of Health Policy & Management and of Health Informatics at the Johns Hopkins Bloomberg School of Public Health, and is director of the Johns Hopkins Center for Population Health Information Technology (CPHIT), which was founded in 2012. The Johns Hopkins CPHIT, under his direction, has been very involved not only in academic research and study around the intersection of population health initiatives and health information technology; it has also been deeply involved in a new initiative designed to boost the all-payer hospital rate regulation system unique to the state of Maryland (the Johns Hopkins organization is based in Baltimore).
As Kaiser Health News reported on January 10,“Maryland officials have reached what analysts say is an unprecedented deal to limit medical spending and abandon decades of expensively paying hospitals for each extra procedure they perform. If the plan works, Maryland hospitals will be financially rewarded for keeping people out of the hospital—a once unimaginable arrangement.”The KHN report went on to note that, “After months of negotiations with state and federal officials, the hospitals also agreed that their revenue from all sources—private insurance, government and employers—will rise no faster than growth in the overall state economy.” And the report quoted Uwe Reinhardt, Ph.D., the renowned Princeton University healthcare economist, as saying, “This is without any question the boldest proposal in the United States in the last half century to grab the problem of cost growth by the horns.”
As a Jan. 31 article in USA Today noted, “A key ingredient of the state's regulatory scheme is a 36-year-old waiver from the federal government that allows Maryland to set its own reimbursement rates for Medicare, instead of following the rates for the country as a whole. This new plan replaces that waiver and requires the state to cut total Medicare costs by $330 million over the next five years. The waiver also requires Maryland to reduce its hospital readmission rate and cut hospital-acquired infections and other preventable conditions by 30 percent.”
Given both the opportunities and challenges inherent in the Maryland payment initiative, there is no question that data and IT will be key elements in the potential success of the program. The state has an all-payer database, and a level of collaboration among providers and between providers and private and public payers, that is rare in the U.S. And that’s where some of the functions of the Johns Hopkins CPHIT come in, notes Dr. Weiner.
Specifically with regard to the statewide payment initiative, Weiner says, “We’re trying to link data from EHRs, vital records, and claims data; we do have an all-payer database here. And the data are everywhere, but they’re just not linked. And to be more specific, at Hopkins, we’re blessed with a lot of doctors, nurses, public health people, engineers, and computer scientists; and our goal is to bring them all together to improve health. “
Dr. Weiner spoke recently with HCI Editor-in-Chief Mark Hagland regarding the Center’s activities broadly and also in relation to the Maryland initiative. Below are excerpts from that interview.
Tell me about the work of the Center for Population Health Information Technology in general.
There are very few people trying to create connections between EHRs [electronic health records] and population health, and public health. Our organization is trying to create those connections, and it’s based in a population health school. The soft start for our center was in late 2012. And our goal is not just to focus on the person with diabetes who’s in front of a doctor, but on all persons with diabetes. From time to time, clinicians and hospitals focus on the broader populations; but that is at the core of our activity.
And how are you accomplishing that work?
Our goal is to tap into the data coming out in the medical environment, through EHRs. Some 85 percent of doctors are already using EHRs. And we’re thinking about how to leverage that data; and we’re focused on public health. But we didn’t call it the center for public health IT, for a number of reasons; we think the trick is to focus on the entire healthcare system. We do studies, we do syntheses of knowledge, we develop tools. In addition, I and my late colleague Barbara Starfield developed something called Johns Hopkins ACGs—adjusted clinical groups (originally ambulatory care groups). In the early 1980s, we were working with diagnostic data and pharmacy data in our own health plans, and we developed tools for predictive modeling, for risk stratification purposes. ACGs work mainly with claims data.
But we’re now using data from EHRs, from mhealth, from vital records, and we’re linking data sources. For example, here in Maryland, on January 1, we started with a new model that includes an all-payer hospital system. Until recently, the payment was based on case type. Now, Medicare, Medicaid, and commercial payers all the pay the same rates in our all-payer system. But as of January 1, it’s moving towards a globally capitated budget based in part on health data.
So part of what you’re doing will help support the new plan in Maryland?
Yes, we’re trying to link data from EHRs, vital records, and claims data, to support that program. Our goal is to bring everyone together—doctors, nurses, public health people, engineers, and computer scientists, to improve health.
What kinds of tools are you bringing into this work?
We’re doing natural language processing and text mining, and we’re doing work here and with a few integrated delivery systems. As you know, a good percentage of the data in EHRs is inaccessible, because it’s in freetext. And we’re working with four integrated delivery systems to tap into predictive modeling, using data, for risk stratification, using new sources of data.
Meanwhile, our school is extremely international; there are more colleagues working globally than in the U.S. And though we’re focused on the U.S., we’re also trying to find some global projects that will help our international colleagues.
Can you share a bit more background regarding the Maryland statewide payment initiative, to help our audience understand the policy context of that initiative?
Certainly. At one point, there were a dozen states in the 1970s that had sort of a public utility rate system for hospitals as an approach to cost containment; Maryland is the only one left. Here in Maryland, they’ve tried to maintain the positive elements of that kind of system—all-payer, fairness to the uninsured—while updating it in terms of the value to the community with population health. So we are, with a very few exceptions like Massachusetts and Vermont, the most progressive state in the nation. And they will be paying providers largely based on community health improvement. They’re looking at all-payer claims data, public health metrics, ambulatory care-sensitive conditions—in other words, are people being hospitalized based on things they could have taken care of in the community?
And by the way, they do have fairly advanced elements in the program, like mandatory reduction of avoidable readmissions. In fact, several rural hospitals have already moved to this; indeed, one rural hospital took over the children’s school lunch program, to support public health. But the challenge with hospitals is, what is the community, if you’ve got six hospitals in a city? And also, urban Maryland crosses into D.C. [the District of Columbia], and western Maryland crosses into Pennsylvania; so there are some challenges. But there’s a complete control over all hospital costs. So they’ve agreed to limit all payers and all hospitals (inpatient and regulated outpatient care); have agreed they cannot go above 3.6 percent inflation per year; and this Medicare, Medicaid, and all private payers. And all uncompensated care is rolled into that, so hospitals are made whole on uncompensated care. And Medicare pays exactly one dollar on every dollar—Medicare, Medicaid, and all private payers are paying the exact same rates; that’s the only state in the union where that is true. The all-payer system has been in place since the 1970s, but moving towards population health metrics and a global cap, those elements began January 1.
I’ve heard different nomenclature here—what is the system called by most people?
Sometimes, it’s called the Maryland all-hospital, all-payer system, sometimes the Maryland waiver system. I call it the all-payer, global hospital payment system. Other people have called it the Triple Aim payment model, related to IHI’s Triple Aim, since it’s focusing on population health, improved care quality, and cost. What’s really unique is some of the population health metrics.
What are some of those metrics?
In the first round, they’re using things like ambulatory-sensitive conditions, so if people are hospitalized for things they shouldn’t be, you lose points; they’re using public health metrics from the health departments. They’re also going to be doing surveys on the patient experience.
And some of the metrics from those surveys will be applied to payment?
Yes. Normally, if your cases go down and the number of patients go down, you normally lose payment. That won’t necessarily happen here; if your costs go down but your quality goes up, you could potentially remain the same. It’s in some ways like turning hospitals into anchors for managed care. Now, it would include certain outpatient services delivered by the hospitals. But it does not include services provided by doctors in their practices. So most Part B services under Medicare will not be included.
So one of the challenges, in fact, is, how do we blend the Medicare primary care medical home practices here? Maryland also has one of the biggest primary care medical home programs in the country, via Blue Cross Blue Shield. CareFirst, which is Blue Cross, is the biggest insurer in the state. They have the majority of commercial patients. And there is no Medicare Advantage in Maryland; but the Part B Blue Cross has a demonstration project, which involves virtually all the primary care doctors in Maryland who serve Medicare patients. They also have an all-payer pilot project, where they’re trying to build medical homes that span the Medicare, Medicaid, and private-payer environments. That’s a state-sponsored pilot project.
And the next thing is that the state is about to submit a proposal for an innovation grant, like the innovation grant in Vermont. If things go as expected, Maryland will get a grant to link all these medical homes, to take the existing infrastructure I’ve just described, and to fill in the pieces, mainly Medicaid, and some of the other commercial plans, and create a statewide primary care network through facilitation; it won’t be run by the government.
And the reason we need this is that our hospital system will arguably be the most structured in the country; and the goal will be coordination across the state.
What will be learned from the innovation grant, if it is approved?
The federal government is going through formal evaluations. And CareFirst, to their benefit, is being very thorough in their preparation. So I think it will be to understand the impact on cost and quality. But back to our population health group at Hopkins, information will be the glue that will hold it all together, in terms of broader population health work. Everybody has some existing data. We have an all-payer claims system. We do have a new, very effective HIE [health information exchange]. And because of state regulations, every hospital is an active participant. In fact, one of the projects… the Center for Population Health IT, or CPHIT—our goal is to help link things together. About 85 percent of all doctors now have EMRs, perhaps slightly higher here than the national average. Are those EMRs all talking to each other? No, not yet. And in fact, the goal of CPHIT is to link all the EMRs of all the physicians and all the hospitals. So we are collaborating with the various partners to help them with technology development. And some of the challenges are standards across these different systems; some come out of, how do we develop metrics, not at the visit level, not at the discharge level, but at the person or community level? And that means you cannot just use one EMR or hospital record; you have to link them all. And that’s not easy.
But in a sentence or two, that’s what CPHIT is all about. And it’s not easy, and to try to chip away at some of these challenges requires bringing together experts from many areas. We had an exciting meeting yesterday with the leadership of the state health department and the commissioner of health.
Will a lot of this be activity replicable in other states nationwide? There’s so much diversity in different healthcare communities and different regions of the U.S.
Your point is well-taken; there are scope and leadership issues. And I think we’re a smart state here, but it wouldn’t happen without financial incentives. And it’s ACOs and other organizations that have to pull things together. And I believe it wouldn’t happen in Maryland without those incentives.
But there’s a broad consensus within Maryland about the importance of moving forward in these areas?
Yes, and because of this waiver, far more money is coming into Maryland hospitals because of the federal payments are higher; and the federal authorities are expecting something in return in terms of cost versus quality, from this. And if the waiver hadn’t been approved in January, by CMS, I’m not sure that this could succeed. But Uwe Reinhardt, who’s not shy, calls this the boldest change in healthcare reimbursement in 50 years. It’s pretty big; so we’ll see.
