The concept of health information exchange (HIE) is far from new; indeed, as those who have been around the healthcare industry long enough can attest, we are already into our third iteration of the concept, with community health information networks (CHINs) first emerging in the early 1990s, followed in the first decade of this century by regional health information organizations (RHIOs). With core problems around governance, local market competition, financial sustainability, and basic information technology plaguing CHINs, most died a relatively quick, though painful death; and most RHIOs faced severe governance and sustainability issues.
In contrast, the number of HIE organizations nationwide has grown dramatically in the United States in last few years, with the federal Agency for Healthcare Research and Quality (AHRQ) estimating that there are currently over 280 HIEs, and that more than 50 percent of the hospitals in the country are participating in HIE organizations. The same article on the AHRQ Web site that cites those figures cites achievements that have already been documented for HIEs. These include “reduction of duplication and operational costs”; “improvement of quality and health outcomes”; “improvement of public health surveillance”; and “strengthening of links between health-related research and actual practice.”
What’s more, virtually all responsible leaders across the U.S. healthcare industry support the concept of health information exchange; and HIE is enshrined in the Health Information Technology for Economic and Clinical Health (HITECH) Act, through the meaningful use process.
Even as HIEs continue to grow in number and in breadth of scope, healthcare leaders within and outside HIE organizations see stumbling block after stumbling block facing the sector in the next few years, and are asking what, if anything, could be done at the federal level to address the problems they see as hampering the long-term success, stability, and sustainability of health information exchange. Among the problem areas they see are:
• The lack of highly granular data-exchange standards—with some in the industry arguing for federal government intervention in that area;
• The broad lack of interoperability between HIE processes and electronic health records (EHRs);
• A combination of financing and governance issues that speak to long-term sustainability problems: many existing HIEs got their start through federal and/or state grants, many of which are now expiring or have expired; and unless consensus-driven strategies can be developed, many HIEs will falter once the grant money runs out;
• Underlying this, a failure to achieve alignment of goals among the stakeholder groups in HIE organizations; and
• A failure on the part of vendors, according to many industry observers, to provide the leading-edge technologies needed to break through EHR interoperability-related and other barriers.
All of these issues are becoming clearer at a time of intensifying need for HIE, in order to support accountable care organizations (ACOs), bundled-payment contracting, patient-centered medical home (PCMH) care models, clinical integration, value-based purchasing, and myriad other strategic goals in U.S. healthcare. In short, say many healthcare IT leaders, health information exchange is at an inflection point of its evolution in the current policy and operational environment.
THE MESSAGE FROM MAINE
f there’s anyone in the U.S. who can speak to where HIE is right now, where it should be, where it’s been, and where it’s going, it might be Devore (Dev) Culver, executive director of the Portland, Maine-based HealthInfoNet, the statewide HIE for Maine, which has all 38 hospitals in that state under contract. Incorporated in January 2006, HealthInfoNet has benefited from clear-sighted vision, broad stakeholder consensus, and clearly articulated need, in a state that, outside of a few cities, is largely rural, and where market competition is less of a factor than in many states.
Not only is HealthInfoNet a beehive of data and information exchange, with 3 million messages transmitted a week, and 84 percent of the state’s 1.3 million people represented (there are actually 1.2 million lives in the HIE, but some are also vacationers); Culver and his colleagues are currently building analytic tools to sit on top of their HIE’s data warehouse.
The key lesson so far in HealthInfoNet’s success? “The single most important early lesson was initially building that trust framework that allows information to be shared across competitors,” says Culver. “The second lesson is to treat this like a business. At the end of the day, if I can’t add tangible value, you’re probably not going to pay me. That’s the root challenge across the country when we look at health information exchanges. So here I am, a not-for-profit organization, but revenue has to exceed expenses, or we’ll go out of business. Over the last three or four years, the concept of an enterprise exchange has come onto the table, where a group of hospitals or providers reach out and build a private exchange.”
Jody Cervenak, a Pittsburgh-based principal with the Denver-based Aspen Advisors consulting firm, agrees, adding that “The reality is that I think that the major underlying obstacle comes down to aligning incentives, because if we align incentives” among stakeholders, “progress on standards and models will take place.” Offered the metaphor of a long aisle of cereal brands at the grocery store with regard to all the different IT infrastructures in HIEs right now, she says, “I love your mentioning the choices in the cereal aisle at the grocery store. You’re right, there are so many types of cereal at the store, right? But they’ve all agreed to put standardized UPC [universal product code] codes on their cereal boxes, for improved efficiency of store management. That standardization was created because everyone in the food industry had aligned incentives: they wanted to get the product to the consumer, faster, cheaper, better. What needs to happen in healthcare,” she says, “is to break down the silos of patient health information and data. And that would mean that my height, weight, age, problem list, allergies, etc., would be presented in some standardized fashion across all the different databases in healthcare. The problem is that we have technology vendors that may not yet have aligned incentives.”
THE CDA AND THE C32: TECHNICAL OBSTACLES TRIPPING UP HIE LEADERS
Meanwhile, interoperability and standardization continue to be a core challenge at the technical-operational level. Tony Gilman, CEO of the Austin-based Texas Health Services Authority, which facilitates health information exchange at the state level, helping the 12 HIEs in Texas share data among each other, puts it this way: “Interoperability and standardization continue to be a problem, particularly as we’re working with providers using products that existed before meaningful use and thus have deployed pre-meaningful use standards. Those products are making progress, but aren’t moving fast enough. ONC and CMS [the federal Office of the National Coordinator for Health IT and the Centers for Medicare & Medicaid Services] are really focusing on the consolidated CDA”—the Clinical Document Architecture, an HL7 standard—“and they’ve focused on that standard for patient summary exchange.”
Federal approval of broad standards is in itself a good thing, Gilman says. “But the problem is that the vendors aren’t there yet, and the systems in place [products] don’t support that. So a lot of them can exchange a C32 document”—the HITSP 32 Summary Document using the HL7 Continuity of Care Document standard— “but that document often doesn’t include problem lists, allergy information, or the level of detail you would expect; a lot of it is just demographic information.” In other words, Gilman notes, there is a level of standardization of approach at the broad policy level, but that has not yet translated into true ground-level interoperability.
In other words, there remains rather a large gap between policy-level and EHR-level adoption of standards across the U.S. healthcare industry. Many HIE leaders agree that the transmission of CDA-level data across organizations continues to stumble when it comes to moving the data in such informational packets into EHRs, because of ongoing differences in display presentation at the individual data field level. Or, as Mark Frisse, M.D., professor of biomedical informatics and director of regional informatics initiatives at Vanderbilt University in Nashville, puts it, “Most EHRs can’t really fully accept documents. They can receive lab results, but when they receive inbound documents, the question arises, whose record do we attach an inbound document to?”
Further, Frisse asks, “How do I make sure Joe Smith’s inbound documents get to Joe’s record? If there’s a request, then I get it in my world. But do I put it in my record or not? Presumably, you could take a whole CDA and throw it in there as a PDF blob, but most institutions have policies that say, it’s not a part of my legal medical record. There are just a million logistics issues on the ground. The problem there is when the Stage 2 regs say that 10 percent of my summaries have to go out in electronic form that means that 10 percent of my referring docs have to receive it in electronic form. People are saying, OK, if you can receive it as a fax, that’s OK”—which means that some of the hoped-for automated patient data flow is still not happening. “People in the policy arena believe that can you solve 30 years of issues around clinical practice with a few conference calls and meetings,” Frisse adds.
CALL IN THE FEDS? THE FEDS THEMSELVES AREN’T SO SURE
Russell P. Branzell, a former CIO who this spring became CEO of the Ann Arbor, Mich.-based College of Healthcare Information Management Executives (CHIME), has been speaking out publicly, including in May at the Healthcare Informatics Executive Summit in San Francisco, about what he sees as the need for federal government intervention in this area. As Branzell, who calls himself a “small-government kind of guy,” sees it, ONC or CMS needs to intervene, soon, to establish data-exchange standards at a far more granular level in order to address the kinds of concerns described above by Mark Frisse. In late May during a visit to Washington, D.C. to meet with federal legislators and agency officials, Branzell reported that “We had a specific discussion with CMS, and the key to HIE success will be standards, standards, standards,” Branzell says. “We’re still working with a level of standards that is not granular enough to eliminate variability; there is still tons of variability at the EHR level. In my view of the world,” he adds, “this is a rare example of where you want the government to be as prescriptive as they possibly can,” in order to help resolve this EHR-level interoperability problem.
But do the feds want to step in? Judy Murphy, R.N., deputy national coordinator for programs and policy, says that everyone should pause to consider what’s being asked. “Russ is coming from the standpoint of ease of implementation,” Murphy says, “and what he’s describing, actually, is constraining the standards, so that we would specify, Use Block A, Use Block B, and Use Block G. Some folks have gathered together to start to do that. For example, there’s a collaborative in New York that is constraining the standards; and then it becomes more plug-and-play, because you’re using a specific standard, and therefore, the implementation is actually easier.” But, she adds quickly, “Basically, I do not believe that there is an appetite in our industry for a lot of specificity in this space. The reason I say that is that there already has been a lot of discussion of this through the Standards & Interoperability Framework, which was launched in January 2011 by the ONC [http://wiki.siframework.org/]. And we have not heard a call for that in that space,” Murphy adds. “In fact, the folks developing standards have wanted to use a building-block approach or the option of being able to pull standards in; and more importantly, using a consensus-based process for building those standards. Unless the appetite of the entire industry changes dramatically in this space, I’m not sure that we’re going to get a lot of support for changing this model.”
Judy Murphy, R.N.
MOVING FORWARD, STEP BY STEP
In the meantime, HIE leaders across the country continue to move forward to build sustainable HIE infrastructures and process. For example, Chris Carmody, president of ClinicalConnect, has been leading a broad regional HIE based in Pittsburgh, one that encompasses nine health systems, three long-term care facilities, one pediatric rehabilitation hospital, and one pediatric physician practice, in western Pennsylvania. In fact, Carmody has a dual role; he continues to serve as vice president of infrastructure at the Pittsburgh-based University of Pittsburgh Medical Center (UPMC), one of the founding member organizations in ClinicalConnect.
Data exchange went live among ClinicalConnect members in June 2012; and in addition to the nine health systems participating, the HIE counts about 8,000 physicians, more than 6,000 of them employed by or affiliated with UPMC; the data exchange at ClinicalConnect already encompasses more than 8 million unique patient records in western Pennsylvania. The biggest challenges? According to Carmody, establishing trust among member organizations that are robust market competitors, and developing a comprehensive HIE on a relatively lean budget. He and his colleagues have been partnering with the Pittsburgh-based dbMotion and the Chicago-based Initiate Systems (now a part of the Armonk, N.Y.-based IBM). Carmody says that “I see a bright future ahead of us, with 10 to 20 more organizations joining us in western Pennsylvania.
On the eastern end of that state, the University of Pennsylvania Health System (Penn Medicine) is participating in the Delaware Valley Health Council, which encompasses 44 hospitals in southeastern Pennsylvania, confirms Michael Restuccia, vice president and CIO of Penn Medicine. At the same time, Restuccia notes, Penn Medicine is using the CareEverywhere data exchange capability built into the core EHR solution from the Verona, Wis.-based Epic Systems Corporation.
Asked what he sees as the most important core strategies for a complex multi-hospital health system like Penn Medicine, Restuccia says, “As an organization, you have to define what data you want to exchange. First and foremost, I personally think we should focus on exchanging the clinical patient data. If you put the patient at the center of the universe, what do you want to be exchanging first? I think it becomes a fairly easy question to answer.” Core lessons learned so far? “Don’t underestimate the complexity of being involved in a regional HIE. It costs money, so there’s funding; there are unnatural alliances, such as between providers and payers—that have to be managed. And there’s a whole staffing issue involved.”
Meanwhile, in Rochester, N.Y., John Glynn, senior vice president and CIO of Unity Health System, has been leading his colleagues in Unity’s participation in the 18 patient care organization-member Greater Rochester RHIO, since 2006. Not only was Unity Health well-positioned to participate in the Greater Rochester RHIO because of its early implementation of EHRs across its continuum; it has had a very strong motivation for HIE participation, because of its involvement in a state grant-funded community diabetes collaborative based on patient-centered, coordinated care models.
An absolutely key learning? “We’ve been successful with our HIE development work because we engaged all our stakeholders early and upfront, and because HIE development has not been an ‘IT project,’ but because it’s been a clinical project, one tightly linked to our broader strategic goals” around care management and community health, Glynn emphasizes.
WHAT HEALTHCARE IT LEADERS SHOULD DO?
Given some of the policy, process, and vendor-market uncertainties, what should CIOs and other healthcare IT leaders in patient organizations be thinking about, as they determine whether and how to pursue HIE opportunities? The Texas Health Services Authority’s Gilman says that patient care organization leaders “really have to understand internally what they’re trying to accomplish in terms of health information exchange. So they need to do an internal process to understand where they are and where they want to go, and how that benefits their communities. For example, in Texas, we have a lot of transient patients who go from one healthcare system to another. And there’s value in competing healthcare systems to exchange information to support their patients.”
ClinicalConnect’s Carmody adds this: “If they’re thinking of starting their own HIE now, it’s probably too late. The market is already flooded with different HIEs popping up across the country. If they haven’t decided whom they’re going to connect with, I would first look at what their governance model/structure is—private, public, or a public/private hybrid form. A lot of HIEs got their start with federal grants and are now struggling. You want to make sure you connect with the right HIE, because another one might go out of business. We’ll see more consolidation going forward. Finally, this isn’t easy; it’s a difficult process, and it requires a commitment to exchange data. Where a lot of key work will happen is in leveraging current EHRs and other clinical information systems to move data from current systems into HIEs.” ◆