Most health information exchange efforts employ a distributed model, with data passed between providers' systems. But a few regions are experimenting with a different approach: the creation of centralized health record banks (HRBs).

Health authorities in Oregon and Washington have received grant funding to establish HRB pilot projects, and Louisville, Ky., is establishing a regional HRB.

Some state health IT officials in Washington state are skeptical the distributed model of building interfaces between provider data silos will work. “To be frank, we don't think that's sustainable or feasible,” says Juan Alaniz, manager of the HRB project for the Washington State Health Care Authority in Olympia. “We think the real change will come from consumers having control over their healthcare information and sharing it with whomever they want to. The health record bank could be the disruptive technology to change the paradigm.”

The state of Washington has allocated $3.4 million for pilot projects in several communities, some involving large hospital systems. The state plans to select the pilots this summer and have the first HRBs in operation by February 2009.

Alaniz says hospitals plan to offer their patients personal health records (PHRs) that tie directly to the HRB. Some may build on work already done in the Bellingham, Wash., region on a Web-based PHR called My Shared Care Plan.

“Rather than build interfaces to other healthcare providers,” Alaniz says, “they can just build one interface to connect to the bank.”

Last year, Oregon got a $5.5 million federal grant to begin development of an HRB for the 400,000 clients of the Oregon Health Plan, the state's Medicaid system.

“Like 99.9 percent of the healthcare system, we are dealing with fragmented data sources, and the information that clinicians need is not available to them as Oregon Health Plan clients move between multiple places for care,” says Greg Wenneson, project manager for the Health Record Bank of Oregon.

Launched in October 2007, the project's concept is to build a data warehouse so that clinicians can access their patients' records via the Web. The patients would retain control over who could see the information and any secondary uses of their data.

The CIOs of several large hospitals are on the project steering committee, and Wenneson describes them as enthusiastic. Technology is not their primary concern, he says, describing privacy and security policies as more problematic. But perhaps the most difficult question is how to make the system patient-centric. With many HIEs, the exchange is between organizations, and the patients are passive. The question is how engaged are the patients going to be. “If the patient does nothing, what do we have?” Wenneson asks.

Perhaps furthest along is Louisville, Ky., where after a two-year research process, a nonprofit consortium is in the final stages of approving a business plan for the Louisville Health Information Exchange (LouHIE), based on an HRB.

LouHIE's business plan calls for selecting a vendor to work with in the fall 2008, piloting a project in the second quarter of 2009, and building out the system in 2010.

The biggest challenge is to get members of the public to sign up for the free records at their doctors' offices, says Judah Thornewill, director of the Center for Collaborative Communities at the University of Louisville and LouHIE's acting executive director. “CIOs and practitioners tell us the health record bank can't have less than 50 percent of patients in it. We need a majority, they say, or it becomes more trouble than it's worth to fit it into their work flow.”

LouHIE will also work on a method to extract the data it has permission to collect from hospital computer systems, Thornewill says. “The hospital CIOs understand how complex this is and they were skeptical that we could achieve this vision without a number of years of work.” They are having a hard enough time just integrating the different levels of technology they have within their own organizations, he adds.

That is why LouHIE's plan is to start with something it perceives as manageable, such as medication data reconciliation, and to work to get 70 to 80 percent of the patients in the area enrolled, then expand the features slowly from there. “What clinicians say they want is Web access to a patient clinical summary, with information on medications and allergies,” he says. “It's a deal-killer if they are not proponents, but if that medication information and clinical summary fits into their work flow and only takes another 20 seconds to do it, we think that will fly.”