Drilling Down on Data Analytics for Value-Based Care Delivery

Oct. 4, 2016
What’s being learned as providers move into the trenches in their ACO and population health work? Industry leaders at the World Health Care Congress noted the gains from early data analytics efforts

What have proven to be some of the major stumbling blocks facing the stakeholders of U.S. healthcare as they pursue data analytics to support value-based care delivery and purchasing? A plethora of issues emerged during a discussion at the World Health Care Congress on Monday in Washington, D.C., during a session entitled “Analyze Delivery System and Payment Reforms to Drive Quality and Control Costs in Government Programs.”

The discussion was led by Enrique Martinez-Vidal, vice president, state policy and technical assistance, at AcademyHealth, a not-for-profit organization that, according to its website, “works to improve health and the performance of the health system by supporting the production and use of evidence to inform policy and practice. Martinez-Vidal was joined by fellow panelists Navneet Kathuria, M.D., M.P.H., vice president, population health and clinical quality, at Meridian Health, a Neptune, N.J.-based accountable care organization (ACO); Wayne Turnage, director of the District of Columbia Department of Health Care Finance (DCHCF); Jeffrey Spight, president of the White Plains, N.Y.-based Collaborative Health Systems consulting group; and Anna Keith, vice president of innovative solutions, at Lifeshare Management Group/the Centene Corporation (Washington, D.C.).

The panelists described a very broad range of challenges when it comes to leveraging data analytics in the new healthcare. Dr. Kathuria of Meridian Health noted that “My activities are focused on managing populations for all our populations. I oversee our system’s data warehouse. The biggest challenge,” he said, “is getting all the data from all these disparate systems, having it churn through our system, and then analyze the output, and partnering with the 1,200 physicians in our commercial network, 800 of whom are in our Medicare ACO. So our infrastructure for data analytics is an ongoing issue and challenge, because data comes in at different speeds. And how we work with our employed and non-employed physicians, is a challenge.”

Spight, noted of his organization, “We work with about 5,000 physicians around the country, on MSO and ACO operations. We have 24 MSSP ACOs and one Next Generation ACO across our portfolio. And everyone’s looking to expand into Medicaid and commercial as well.” And he said that, “For us, the biggest challenge was getting the right staff. We thought we could take people with medicare Advantage experience. That just didn’t work that way. We needed people who were problem solvers. And really, it was, how do you change culture, how do you change behavior?”

Turnage reported that, “From the government side, the environment has been favorable. We have not received any pushback from the District side or CMS [the federal Centers for Medicare & Medicaid Services] around innovation. As far as the policy, everybody’s on the same page.” That having been said, he quickly added, “The challenge in terms of the provider community is, how far do you go in leveraging risk? For me,” he said, “the most effective way to drive outcomes is for the provider to be entirely responsible for the cost of patient outcomes, still allowing them to make some money.” But that ideal ends up bumping against the reality that providers cannot always sustain a level of performance that will support any fixed payments that might come from agencies like the DHCF, he said, meaning, that they can easily find themselves in untenable financial positions, thus endangering the risk contracts they might sign in the first place.

With regard to the kinds of capabilities needed in these areas, Anna Keith noted that a year ago, the Washington, D.C.-based Centene Corporation had acquired the Manchester, N.H.-based LifeShare Management Group. “They purchased the company for its I/DD [intellectual and developmental disabilities] solution, so that as it went into markets that were seeking Medicaid I/DD managed care operations, we brought that expertise to the table. We also brought people services management to the table.” She noted that complex sets of capabilities are going to be needed to master population health and accountable care challenges going forward.

Learnings out of challenges

Still, the panelists noted, a lot has been learned from early data analytics work, despite the initial challenges everyone has faced. For example, Dr. Kathuria noted, “My organization saw the MSSP [participation in the federal Medicare Shared Savings Program] as learning to ride a bike with training wheels, and that’s what we’ve done. These were community hospitals where the majority of physicians are independent contractors, and the MSSP was a way for them to participate, in a non-threatening environment,” he said. “As a result, we realized there were a lot of gaps” in a variety of areas of performance, he reported.

What happened, Kathuria said, is that he and his colleagues at Meridian Health quickly realized that they could not make the progress needed, working across a combination of 40 different electronic health records, and with 40 percent of primary care physicians still working on paper. As a result, he said, “We’ve invested in a centralized data warehouse. And we realized we could no longer work with 40 different EHRs, so we’re moving into Epic. And,” he added, “we realized we didn’t have true care coordination, so now we’re engaged in looking at care coordination across our continuum and having a care coordination IT system.” Further, he said, “It’s as a result only of what we’ve learned in the last three or four years that we’re able to go into some commercial pay-for-performance contracts. Investing in infrastructure came out of that,” he emphasized.

Anna Keith reported that she and her colleagues learned a lot from developing early population health programs, including the need to tie care management efforts closely to the day-to-day lived experiences of patients under care management programs. For example, she said, “How much do you think that someone who is living in his car is really going to care about the pieces of paper you hand them about smoking cessation and weight loss?” Data analytics, she emphasized, can help the leaders of patient care organizations look at broad populations, but the results that come out of such analytics work will always need to then be translated into efforts that can work within the parameters of the day-to-day lives of the patients and health plan members involved. In other words, care management plans in the real world inevitably need to be tailored to the lived experiences of patients and plan members.

Speaking of reality-based operations, consultant Spight said, “We should keep in mind that the idea of interoperability is a unicorn at best. And we end up spending time worrying not about what’s in the EHR, but instead, how we get the data that isn’t in the EHR, including real-time lab values, hospital discharge and transfer files, and other pieces of information, and get that information to physicians.”

Still, even early efforts using data for population health and accountable care purposes are already being applied to forward-facing efforts in the patient care organizations that have done some data analytics and population health work already. As Kathuria noted, “All of our hospitals are in the CJR”—the federal Comprehensive Joint Replacement Model that holds Medicare-participating hospitals responsible for the outcomes and costs of comprehensive joint replacement procedures. Based on early data analytics work for their population health and ACO participation, he said, “We were able to get all our orthopedists in the room, because they had been listening to the value-based purchasing discussions, even though they hadn’t been engaged in our value-based purchasing work. So the work that has happened within our organization with regards to the MSSP has allowed us to develop at least the knowledge and the language to engage.”

And the work continues, Kathuria noted. “It's not enough to have a data warehouse, you need to be able to mine it. You need individuals who understand clinical informatics, as well as big data analytics. And nothing is turnkey. You’ve got to develop the interfaces. You’ve got to make sure that what you call a cardiologist is what the payers call a cardiologist. You may have a cognitive cardiologist, an interventional cardiologist… We have five different hospitals, and each was using a different labor for glycolated hemoglobin. And we had different definitions. So you had to make sure the variables were made identical. So it’s one thing to buy something off the shelf, it’s another to be able to use it. Even the vendors are struggling. So you have to have a little bit of knowledge and expertise in-house.”