No one said population health management would be easy, nor has it been. Indeed, as patient care organization leaders have laid the foundations for serious population health initiatives, they’ve been facing down a welter of challenges, among them strategic, process, practical, and technological. But things are beginning to turn a corner now.
And the truth of that perception is confirmed by the results of a survey published by consulting firm KPMG in January. According to that survey, provider and health plan leaders are making progress in key areas. As stated in the consulting firm’s Jan. 23 press release, “In the survey, 44 percent of respondents at payer and provider organizations found that they have a population health platform in place that is being ‘utilized efficiently and effectively.’ Another 24 percent are in the process of implementing a population health program within the next three years. Only 10 percent said they have no plans to implement a population health platform and another 21 percent of respondents said their organization doesn’t require a population health platform.”
Importantly, KPMG found that “The biggest individual barrier to implementing a population health program is aggregating and standardizing information from multiple sources, 30 percent of respondents said. Stakeholder adoption (10 percent) and integrating with clinical work flows (10 percent) were cited as additional barriers. Another 34 percent cited ‘all of the above,’ which includes those barriers, as well as enabling patient engagement, funding investments, and selecting appropriate vendors as additional challenges.”
Gathering data from diverse sources remains a core challenge
Jess Vamvas, senior product manager, technology, at The Advisory Board Company (Washington, D.C.), is not surprised by such results. “Apart from the overall strategic planning, the analytics is the most complex part of this,” she says, referring to population health management initiatives. “Organizations want to assemble high-performance networks; and along with that comes the various different source information systems and applications the different organizations are already suing. So getting the data from the EMR is really complex; same with combining clinical and claims, even though that’s where everything is moving towards. Even professional billing services data is really, really important in order to aggregate and get a comprehensive view for population health. For instance, registries have long been important, but now, pre-registries—who could become a diabetic, not just who already is.”
Dennis Weaver, M.D., executive vice president and chief medical officer at The Advisory Board Company, emphasizes that there are several layers to the complexity involved in leveraging data and analytics to support population health. “We can dive down into the clinical examples of the registries and the other things going on,” he says, “but we also need to talk about these multiple stakeholders and the financial aspects of population health management. Too many organizations look at people, process and technology—and they look at these tools that they need to buy for pop health, and then on an economic side of the ledger, they’re looking at accountable payment, whether pay for performance or shared savings, or any number of arrangements. And too many folks just do it based on those payments.”
In other words, Dr. Weaver says, most patient care organization leaders are sabotaging themselves by looking too narrowly at the implications of the population health management framework for their overall operations. And that really speaks to the core strategic challenges of taking on financial risk to begin with. “When we look at population health management,” he says, “we find that focusing on the payment misses two other areas of return beyond accountable payment. As you start to take care of defined populations, the economic return turns out to be 7 to 10 times greater if you can keep that population inside your CIN [clinically integrated network] or your ACO [accountable care organization]. Some people call it reduction of leakage, domestic utilization, or seepage. But the economics here and the quality return that you get if you can keep the patient inside the management system in which you’re doing population health, are very important.”
Dr. Weaver goes on to say that “The third piece of this is an element that people often don’t see as connected to population health; but that is that we’ve got much unwarranted clinical practice variability—and we don’t spend enough time talking about reducing that variability. Because doing that reduces cost and improves quality. There are the clinical components of pop health, but the economic components are so important, too.”
And all that, he says, makes population health “so complicated. And when you talk about the IT components, there’s the clinical, there’s the financial, there’s the patient-facing. But too many folks try to get the ROI just from the payment piece, but not from the network management and clinical variability pieces.”
A complex—and sometimes-messy—ongoing evolution
The fact is that none of these issues exist in a vacuum. Instead, as Joe Damore, vice president of population health management at the Charlotte-based Premier Inc. notes, IT and data analytics infrastructures to support population health work have meandered forward at most organizations along with their increasing involvement over time in different types of risk-based contracting. “A large organization may have put in an ACO for their employees, may have several Medicare Advantage contracts, or formed their own Medicare Advantage plan on their own or with a major payer; they may be in the Medicare MSSP or Pioneer or Next Generation program; they might have several ACOs with commercial payers, and also may have several bundles going,” he notes. “And those activities represent those of an advanced organization, where they’ve been creating alignments. Meanwhile, on the delivery side, they would have changed the metrics and compensation models for their physicians to match the metrics of their models.”
In that context, Damore says, “shifting from an RVU model to a hybrid model that includes metrics that match those in their contracts,” naturally will create “alignment around metrics,” and shift the IT and data needs for such an organization would naturally shift as well. And, he says, “Once they develop a clinically integrated network that would allow for single-signature contracting with all the payers, one with hundreds of independent physicians, but they could reward those independent physicians with shared savings based on the metrics on those contracts, then the independent physicians would be working off the same metrics. Also then, they would have in place a care management program that would focus on at least three areas.”
In all this, Damore says, “The first piece of IT you need is a claims analytics tool. You’ll need to take the claims from Medicare or the private payer, and identify the people at risk, especially those with a chronic disease. And then a predictive modeling piece of that to predict which people will rise up—the rising-risk people.” Unfortunately, “I would say that fewer than 10 percent of U.S. patient care organizations” have those elements—both a claims analytics tool, and a predictive modeling capability to predict which covered patients will become higher-risk over time, says Angela Lanning, the ITS (Information Technology Services) COO at Premier. Those patient care organizations that have already logged experience with Medicare Advantage work “are generally a bit more advanced,” she says. Still, Lanning adds, “The EHR [electronic health record] vendors claim they have the core capabilities to do this, but that’s not really the case. And a lot of times, the physicians in these rural areas—they don’t have the infrastructure to support this,” she says. “We see that even with Medicare quality data reporting.”
Executing on strategy in Southern California
Patient care organization leaders who are in the trenches right now on population health are finding themselves working their way forward along multiple dimensions at once, as those industry experts have noted. In Orange County, California, what that looks like is inevitably complicated, says Kevin Manemann, president and CEO of St. Joseph Heritage Healthcare, in the Los Angeles suburb of Anaheim. St. Joseph Heritage Health Care is a statewide physician organization within St. Joseph Health, which encompasses 2,000 physicians, 800 of whom work within a medical group structure, and 1,200 who practice within an IPA (independent practice association) structure. Meanwhile, St. Joseph Hoag Health, the hospital organization, encompasses seven hospitals in Orange County. St. Joseph Hoag Health is a division of the statewide St. Joseph Health system, the umbrella organization for both St. Joseph Heritage Healthcare and St. Joseph Hoag Health. (In addition, St. Joseph Health and the Seattle-based Providence Health recently came together into a combined organization, St. Joseph Providence Health.)
St. Joseph Heritage Healthcare not only is in the Medicare Shared Savings Program (MSSP)—MSSP 1, to be specific—in addition, Manemann says, “We’ve formed a clinically integrated network, and do some work with the PPO in that business. And we’re moving more towards risk contracts.” In that context, he says, “One challenge we face here is that one out of every two Medicare patients in Orange County is in Medicare Advantage. So all of the MSSP 2 and 3 groups are working with a [per member per month] benchmark [payment] that was established based on your part of the country [the East Coast and Midwest]. In our part of the country, it’s pretty low, compared to Houston or Florida, for example. Houston’s monthly is $1,330 PMPM, whereas our starting point was $800. That’s why we decided not to go with Pioneer”—the Medicare Pioneer ACO program. Instead, St. Joseph Heritage Health joined the regular MSSP in 2016, and has taken on risk for 50,000 lives, while it has commercial risk for 200,000 commercial lives, through United HealthGroup, Anthem, Blue Shield, Cigna, and Aetna.
In terms of working with the physicians involved in risk contracting with the organization, Manemann says that “The strategies are pretty straightforward; a lot of it is around data analytics. We’re using the delegated model, the TPA [third-party administrator] model, more or less, for the risk business. So we’re able to do a lot of risk stratification and predictive modeling. And you have to get your arms around the high-risk patients. So we have analytics that help us identify them, and then we have a program called CareConnect—it involves an R.N., social worker, and pharmacist team that helps manage those patients. That helps keep them out of the hospital and ED.”
Already, 2,500 Medicare patients have been enrolled in CareConnect, or 8-10 percent of the organization’s total at-risk population, including also Medicare fee-for-service. “The high-risk patients represent about 5 percent of our MSSP covered lives, and about 5 percent of our fee-for-service Medicare patients,” Manemann explains. And, the reality, he says, is that “We’ve reached the point where it’s less about analytics and more about use. The bulk of the effort is really about knowing how to lead change and drive change throughout your organization. The bulk of the work now is really around behavior change. There are physicians who will argue the data. And we can pull the data out, and if you have a couple of outliers in a panel of a couple of thousand patients, the direction will be right overall.” Still, with regard to presenting physicians with their data, lags in data availability have a major impact on the dialogue, he says. “The lag can run 60 to 90 days with the lagging indicators, including bed days per 1,000, admissions per 1,000, and ED visits per 1,000; and you can’t manage patients using those lagging indicators. Instead,” he says, “you have to get into your leading indicators—authorizations, referrals, and ED visits—and get those types of data to the clinicians in real time,” to make the care management difference between success and failure in ACO and population health work.
What healthcare IT leaders need to do next
So, what do healthcare IT leaders need to do right now to move forward in this very complex area? What must be done next is “highly variable,” depending on where an individual organization already is, Premier’s Lanning emphasizes. “First,” she says, “you have to understand what data you have now in your systems, to figure out which systems to implement. Even if they have an EHR or a physician practice system, the data required for the measures may be too sophisticated for them. One of the big issues is that in many cases, decisions have been made by [patient care organization leaders] without [their] understanding their data. They’ve really got to step back and understand what data they do have, what they’re missing, and what they need to capture. The problem is that in some cases, the data fields [within different information systems] are empty, or they have to upgrade their EHRs, and that becomes expensive and time-consuming. It’s very expensive, and highly variable. We still have practices that manually abstract data to get the information you need to calculate measures. And that still happens with HEDIS measures. So there’s still a lot of wasteful effort.”
As for what might happen in the next couple of years, Lanning says, “I think we’ll evolve” as a healthcare system. “If you’ll recall, when value-based purchasing started, we had to extract data manually on the hospital side; that’s still true in some cases. The hospitals will continue to buy physician practices. That will accelerate things. I don’t think things will drastically change in two years; maybe in five to 10.”
“I agree,” Damore says. “I think it’s going to be a real challenge for a while. I don’t think we really have the perfect systems today, but we’ve come a long way. In 2001 and 2002, I was using manual chart abstraction to pay incentives to physicians based on HEDIS measures. We pulled samplings of charts, because we couldn’t afford to do all the charts. So the EHR was a major step forward. And we don’t have interoperability yet.”
Still, in the end, Damore says, things are being forced forward on a broad policy level, by the purchasers and payers of healthcare in the U.S. “One factor influencing physicians is that the dollars are going to be more significant,” he notes. “In 2022, someone could earn a 9-percent reduction up to an 18-percent bonus—that’s a 27-percennt swing. The measurement for that would be 2020. So you’ve got two-and-a-half years to get ready for a much higher-risk payment swing under MIPS”—the Merit-based Incentive Payment Program under the MACRA (Medicare Access and CHIP Reauthorization Act of 2015) law. Inevitably, that will mean an intensification of activity to better leverage data analytics to participate in all types of value-based purchasing contracts. In other words, stay tuned: the saga will be morphing forward.