Three months ago, the departments of Defense and Veterans Affairs certified to Congress that their electronic health records could finally exchange data in a meaningful way. But lawmakers aren’t satisfied with that assertion and are looking for more clarity on what “interoperability” actually means.
With some fanfare, the two departments bothattested that they’d met a 2014 congressional mandate to make all of their respective health data interoperable with one another’s IT systems. The claim was based on DoD and VA’s successful implementation of the Joint Legacy Viewer, a read-only web application whose main function is to let VA doctors and claims administrators see a veteran’s DoD records. As of this month, JLV has nearly 200,000 users.
But several Senate appropriators said Wednesday that the functionality JLV offers falls far short of Congress’ original intent.
“Webster’s definition is the ability of a system to work with and use another system,” said Sen. Mark Kirk (R-Ill.), the chairman of the Senate appropriations subcommittee on military construction and veterans affairs. “JLV is a kludge, and it’s a band-aid. It does not provide images, for example. So we’re going to tell our veterans, ‘Welcome to the VA, but we don’t have any of the X-rays the Army or Navy or Air Force did for you while you were in the military. I think most of our members would say that’s not interoperability, and I think you can expect some more definitions of interoperability in legislation from this committee.”
Aside from the fact that the current version of JLV is mainly text-based, Kirk said the system is also inadequate in that it lacks any ability to perform data analytics on the vast trove of digital health information the two departments hold and are now sharing with one another. Those types of analytics could be used, for example, to identify soldiers and veterans who may be at risk of suicide.
“One of my constituents committed suicide after having called into a VA hotline that didn’t have enough people to staff it,” he said. “Companies like Cerner have algorithms that can predict suicide likelihood, but JLV cannot do analytics like that. That is a critical capability we do not have.”
Congress also sees a more robust exchange of interoperable data as one way to speed up the processing of veterans’ disability claims. Although about 11,000 workers in the Veterans Benefits Administration now have access to JLV to help them make decisions on disability ratings, Sen. John Tester (D-Mont.), the committee’s ranking member said the system is inadequate to that task if it only offers a partial view of a veteran’s military health history.
“What we’re trying to solve here is to not have to rewrite the book every time a claim comes up,” he said. “How a person was hurt, when they were hurt, the X-rays, the CAT scans, all of that should be on the veteran’s record so it’s not like a complicated math problem to find all of their documentation. And it’s interesting that VA thinks the text is more important than the pictures. I’m not a doctor, but do you ever do a surgery and not look at the images? Do you just start cutting based on notes?”
Certainly not, said Dr. Jonathan Nebeker, the deputy chief medical information officer in the Veterans Health Administration.
“In edge cases and a lot of surgeries, the images are critical,” he said. “But I’m a geriatrician and primary care provider. For day-to-day care, most of us rely on radiologists to interpret those images.”
And those interpretations are already part of the data DoD and VA are sharing through JLV as one of two dozen “domains” of health care information that follow national standards and that the two departments have deemed to be most relevant for most of the encounters between medical staff and veterans or service members.
Sharing of imagery like CT scans has been held back because of bandwidth concerns, but VA officials said they intend start adding those images to the viewer in September as part of a JLV upgrade.
“The size of the data involved in sharing images is exponentially larger than the radiology reports that describe and diagnose those images,” said David Walkman, a senior IT adviser to VA’s undersecretary for health. “Those images are certainly part of the medical record, but the written report is typically what other providers would use to address their findings and follow a course of care. That said, we’re in the process of delivering an image viewer component. The challenge is to make sure we have the bandwidth to exchange the images when they really are needed for clinical purposes.”
VA is also working on a follow-on project to JLV, known as the Enterprise Health Management Platform (eHMP), expected to be up and running midway through 2018.
Waltman said it would build on JLV, taking it beyond a simple viewer of static records and letting DoD and VA clinicians add their own notes, enter new orders for medications and view those images, regardless of which department’s radiologists first examined them.
“What we need is an integrated capability that brings together all of the clinical data, the process management for managing clinical workflows, integrated with analytics and algorithms that can predict what interventions should be taken and what our care pathways should be,” he said. “We will have the ability to incorporate all off those things by midway through 2018, but I can’t say that we’ll be taking advantage of every use case for all of that data.”
eHMP will replace the Computerized Patient Record System, the software module VA clinicians have been using to read and write patient records since the mid-1990s.
CPRS, in turn, is part of VA’s larger electronic health record system, VistA, which the department has strongly signaled that it intends to to eventually abandon and replace with a new health record infrastructure that it hopes will last for the next 25 years.
That forthcoming system, called, for now, the Digital Health Platform (DHP), is still in the early planning stages.
“The reality is there are new capabilities out there that we need to add. We’re not able to move fast enough under VistA, and that’s why we’re talking about a new platform,” said LaVerne Council, VA’s chief information officer and assistant secretary for information and technology.
Council said the DHP woud be an agile “system of systems” that relies on open source technology and shuns proprietary data standards. She said VA will also apply a new approach to data management, taking into consideration all of the information VA houses on veterans — not just electronic health records — and replacing an estimated 138 regional variations of VistA with a single, centrally-managed IT structure.
“There’s much more to the veteran than health care,” she said. “It’s their benefits, it’s their education, their ability to use our national cemetery system, and we have to do a much better ability of creating that seamlessness. Our intent is to create one capability solution, and everybody has to converge around it. Having 130 different versions makes VistA slow and cumbersome and inconsistent. Moving to an open architecture that allows APIs to come in and use that information and share it and get it back out and share it in a much more seamless area is what we want to do with DHP.”
Because DoD and VA decided to part ways on a previous effort to build a single, shared electronic health record system three years ago, DHP would still be a VA-only system, even if it’s capable of ingesting, without any significant hiccups, the data churned out by the separate commercial health record system DoD eventually bought as part of its Defense Healthcare Management Systems (DHMSM) project.
“VA’s modernization plans raise concerns about duplication with DoD’s system acquisition,” said Valerie Melvin, GAO’s director for information management and technology resources. “Studies have identified 10 areas in which they have common health care business needs and 97 percent of inpatient requirements that are common to both departments. Despite our recommendation to do so, VA has yet to substantiate its claim that modernizing VistA, together with DoD acquiring a new system, can be achieved faster and at less cost than a single, joint system. Uncertainty and important questions remain about what the department intends to accomplish, in what time frames, and at what costs.”