When it comes to value-based care and population health management in the context of health information exchange, it turns out that federated HIE models with non-standardized data inputs simply don’t deliver when it comes to leveraging HIE for data analytics. That was the consensus of a group of healthcare leaders who participated in the panel entitled “Health Information Exchange: New Models of Interoperability to Support Value-Based Care,” on Wednesday, May 14, at the Hyatt Regency in downtown Boston, at the Boston Health IT Summit, sponsored by the Institute for Health Technology Transformation (iHT2; since December 2013, iHT2 has been in partnership with Healthcare Informatics, through its parent company, the Vendome Group LLC).
The panel, which encompassed healthcare leaders from a variety of organizations and backgrounds, was moderated by Judy Hanover, research director at the research and consulting firm IDC Health Insights. As Devore (Dev) Culver put it early on in the discussion, speaking of the struggle that many HIE organizations are having right now to survive and thrive, “The ‘post office’ function of the HIE is probably not a sustainable model. So you really need to be thinking about the data itself. First,” said Culver, who is executive director of HealthInfoNet, the statewide HIE in Maine, “you need to be standardizing the data, to take out some of the variation. In Maine, just about 90 percent of all our residents have some data in the exchange. So we have a lot of data.” And working internally to standardize data, and externally to create a clinical analytics program to leverage data from electronic health records (EHRs), are two elements of what Culver and his colleagues have been doing to advance the population health management capability of his HIE organization. But without having started out with a centralized data governance model, he told the audience, he and his colleagues at HealthInfoNet would never have gotten to where they needed to get, with their HIE.
Panelists discuss HIE, value-based care, and governance
Vish Anantraman, M.D., M.S., director of information architecture and interoperability at the North Shore LIJ Health System, Great Neck, N.Y., noted that “We’re using the HIE as a real-time analytics platform. We wanted to get into bundled payment, and unlike the traditional care management around disease management,” Anantraman said, success in curbing costs under bundled-payment scenarios “comes from connecting these patients to care managers within 48 hours.” And that requires purpose-built architecture and strategy for any HIE, he added.
Bill Young, CIO of Berkshire Health Systems, based in Pittsfield, Mass., noted that “We started with a centralized model.” Even so, he emphasized, “don’t underestimate the energy that goes into standardizing the data. We’ve been putting an intense amount of energy into this, and I’m blessed to have doctors involved, and they’ve been working really hard to get the noise out,” when it comes to creating a standardized approach to terminology, definitions, and concepts, in order to be able to analyze clinical data.
And Heather Nelson, senior director of enterprise clinical applications and IT operations at the Springfield, Mass.-based Baystate Health, and senior director of IT at the Pioneer Valley Information Exchange, noted that “Governance plays a very important role” in making data analysis possible. “Our trading partners have been involved from the very beginning,” she emphasized, “and we’ve spent several months going over the centralized versus federated model pros and cons with them. A few of our trading partners have not wanted to share the data, but their peers around the table have seen the value of having the data centralized in order to perform analytics. So I think governance plays a big role in that.”
Essentially, all of these leaders agreed that getting to the point at which a patient care organization’s leaders can actually perform meaningful analytics on clinical data delivered through health information exchange requires several different elements, including work towards standardization of data elements; and further, that such work is significantly impeded by federated HIE models. It will be very interesting to see how governance models evolve going forward in the HIE world, with issues of sustainability beyond the termination of federal and state funding looming nationwide.