Among the Obama administration's efforts to improve healthcare in the United States, health information exchange (HIE) ranks among the least controversial. HIE lacks the political and public divisiveness of healthcare reform, the government "giveaway" accusations of the electronic health record (EHR) incentive program or the widespread Republican opposition of state health insurance exchanges.
There are two practical reasons why HIE avoids the flashy headlines. When the HITECH Act of 2009 set aside funding for each U.S. state and territory to establish an HIE entity, many private HIEs linking facilities within a large network or specific geographic area were already in place and largely successful. In addition, it's almost universally accepted that patient care improves, and healthcare costs go down, when physicians can readily access patient data at the point of care.
That said, a complex blend of technology, management and financial challenges continue to stymie the further growth of HIE. Unfortunately, solutions aren't easy to find.
Providers Using EHR, Not Necessarily Sharing Them
At the recent Healthcare Information and Management Systems Society's HIMSS13 conference, the Office of the National Coordinator for Health IT issued a request for information on how to "accelerate and advance" health information exchange and interoperability beyond what existing federal policy has accomplished so far.
In the RFI, the ONC notes that EHR use grew 80 percent between 2009 and 2012. While that is good news, data exchange is progressing more slowly. For example, as of 2011, only 25 percent of hospitals could exchange medication lists and clinical summaries with providers outside their network and only 31 percent of physicians were exchanging clinical summaries with other providers. Equally frustrating for ONC is the fact that fewer than one in five patients were given online access to their medical data.
It's no surprise, then, that HIE organizations—the vast majority of which are independent, nonprofit organizations—face numerous obstacles. A recent survey (PDF) by the eHealth Initiative found HIE organizations most concerned about developing a sustainable business model, securing funding, engaging with a wide range of stakeholders (which includes hospitals, laboratories, insurers and other care providers) and convincing those stakeholders that data exchange is secure. HIE entities also face competition from vendors, large integrated delivery networks and other HIEs, all of whom could promise healthcare providers better data exchange at a lower cost.
These challenges are further compounded by healthcare reform, which emphasize two means of delivering care—the accountable care organization (ACO) and the patient-centered medical home (PCMH)—that the eHealth Initiative says can succeed only with "a robust and interoperable HIE infrastructure that can support coordination across the care continuum, data exchange between disparate sources, and evidence-based practices and clinical guidelines for care."
Next-Generation HIE Should Behave Like Facebook
For that to happen, the industry needs what HIE leaders are calling HIE 2.0. This suggests that data exchange needs to move beyond secure messaging and patient data look-up in order to succeed—and, under stage 2 of meaningful use, the government program to encourage EHR adoption, it must. In order to attest for meaningful use and receive government incentives, healthcare providers have to demonstrate that they are exchanging patient records and incorporating that structured data into an EHR.
According to the eHealth Initiative, most HIE organizations are, in fact, doing that. Of the 88 responding HIEs that described themselves as operating, sustaining or innovating—collectively, in eHealth Initiative parlance, "advanced" HIEs—more than 80 percent are exchanging laboratory test results, inpatient data, outpatient data and care record summaries. Fewer than one in three, though, are sharing public health reports, which is a key meaningful use measure.
There's more to be done in other areas, too. Joel Ryba, COO of the Health Information Xchange of New York based in the Hudson River Valley, has identified 16 rules for effective HIE, which he deems the "minimum acceptable level" of functionality that an HIE vendor should have if it's trying to sell to HIXNY.
These rules describe the basics of how data should be delivered, abstracted, located, consolidated, aggregated and made available for auditing and clinical reporting purposes. In addition, Ryba notes, an HIE should include a provider directory, maintain data integrity, provide role-based access control, monitor users' adherence to business rules and support interoperability with native EHR environments.
Few products hit the mark, Ryba admits. As a result, the organization has built a lot of its own health data exchange technology, taking a SOA approach. Before joining HIXNY, he built a similar information exchange for the criminal justice system. The concept is similar, he says—putting disparate records in a single XML resource. "You're just creating person-centric histories from multiple sources," he says.
The key is providing abstraction, which HIXNY does with a loosely coupled common information model. This helps in two ways. One, the data model isn't tied to the outbound consumer of the data. Two, it reduces the number of "adapters" necessary to foment data exchange. With, say, 20 hospitals using 50 different EHR systems, a tightly coupled data model would need 1,000 adapters to exchange data. Loosely couple the data, though, and you need only 50 adapters—which means your nonprofit HIE organization can spend less time building new adapters and more time maintaining the ones it has, Ryba says.
In this sense, Ryba equates HIXNY's work not with HIE 2.0 per se but, rather, with a third generation of electronic HIE. The first pushed data to consumers; Ryba likens it to email. The second let consumers pull data and subscribe to notifications; this mirrors the functionality of websites and search engines. The third generation brings push and pull together and adds secure messaging; this is like Facebook, he says.
50 States, 50 HIE Data Models
Improving HIE to the point that it's as easy as logging onto Facebook won't easy, given the number of connections that an individual organization must make and the challenge of viewing that often-proprietary information once it arrives.
There's no one-size-fits-all solution, either. Five state HIE efforts examined in a recent report by NORC, an independent research organization affiliated with the University of Chicago, present different approaches to setting up an HIE infrastructure.
Maine and Nebraska use a centralized model, in which participants push data to a central repository; in both cases, this statewide infrastructure predates the HITECH Act of 2009. Washington and Wisconsin, however, uses a more decentralized data model, which the NORC report says is typical for states where private HIE entities (often found in urban areas with large hospital systems) were already in place. Texas, with its mix of private HIEs and large, underserved rural areas, has opted to support the existing HIE entities and focus on adding services only where they are lacking.
A centralized approach, which requires significant investments in IT infrastructure, can be more expensive and take longer to implement than a more lightweight data model, which needs only messaging capabilities and provider directories, NORC notes. On the other hand, a centralized data repository makes for better data analysis, which is important in light of federal care quality and public health reporting mandates.
Plus, Ryba adds, organizations can use these rich data sets to "orchestrate higher-level business services." One example is a single, consolidated version of a patient's Continuity of Care Document, of which HIXNY now has more than 1.4 million CCDs on hand.
Another example comes from Vermont, where state healthcare reform places an emphasis in part on the patient-centered medical home; here the central data repository gives all physicians caring for a patient access to his or her EHR data.
An HIE Without Stakeholders Is Unsustainable
Providing such business services is impossible, though, if an HIE entity is unsustainable. According to the eHealth Initiative survey, a majority of advanced HIE entities get more than half of their funding from a single source, though the organizations also remain bullish about remaining sustainable and operational in three years' time, when HITECH Act funding for HIE initiatives will run out. Meanwhile, each of the five states that NORC studied was "universally concerned about sustainability, especially in a rapidly evolving market."
Besides government grants, the most likely funding source for an HIE entity is membership fees. For that to work, the eHealth Initiative notes, an organization must engage multiple stakeholders. This also makes good business sense, since a wider variety of participants means more data and a better picture of a patient's overall health.
This is easier said than done. The ONC RFI on how to improve HIE noted that data exchange was particularly limited among three types of providers: Long-term care, rehabilitation and psychiatric hospitals. Rural healthcare providers, including hospitals and physician practices, also tend to be on the outside of HIE looking in.
Rural hospitals have small IT staffs, not to mention small operating margins, Kolkman says—and healthcare reform efforts to keep people out of hospitals has the adverse effect here of making those margins even smaller. Long-term care is often associated with acute care, which has a reputation for being low-tech, she adds. Finally, data exchange with behavioral health is stigmatized by HIPAA, which many misinterpret to mean that behavioral information such as a patient's medication regimen receives special privacy protection. (It does not.)
Collaboration, then, is key, Kolkman says. Find common ground among all HIE participants. This can be determined by asking what providers want, need and value, as well as what they are willing to pay. This also helps stakeholders identify their "trading partners," or the organizations with which they share a patient population and therefore should be sharing data with.
As with determining a data-sharing model, finding low-hanging fruit is different for each HIE entity. Kolkman recommends running the organization like a business but "providing services as though your life depends on it," namely by adding services only when demand can be proven and promoting those services only when there's a critical mass of data in the system. (The Indiana Health Information Exchange, for example, found that sending emergency room admission alerts and discharge summaries to primary care physicians helped PCPs reduce the number of subsequent ER visits, or readmissions, that their patients made.) If providers continue to look in the data repository but don't see anything, Kolkman says, eventually they'll just stop looking.
HIE 'Sea Change' On the Horizon
Though challenges abound, most experts remain upbeat about the future of HIE.
One reason is the government's commitment to the service, most recently reflected in the releases of an HIE toolkit for rural healthcare providers and three tactical data briefs that provide high-value HIE use cases.
Another cause for optimism is growing interest in the services themselves. Vermont, thanks to a state innovation model (SIM) grant, plans to add data analytics and warehousing, event notification and disease management to its HIE offerings. HIXNY, for its part, is aiming to extend its SOA framework to include business process management and business activity monitoring. "It took me a lot to get the HIE to where it is. I'm not regressing," Ryba says.
The next year-and-a-half should be interesting, says Kolkman, who expects a "sea change" as federal money for the state designated HIE entities comes to an end, the market consolidates as a result and organizations expand partnerships--especially when it comes to sharing HIE infrastructure. This is already happening on the plains, the NORC study says, where Nebraska is selling use of its infrastructure to Wyoming and may expand to other neighboring states.
"We are at an exciting time," says Prashila Dullabh, NORC's program area director for health IT. "EHR is being adopted, the infrastructure's there and we definitely feel we are moving forward."