Last week, an information technology industry advisory group recommended the Veterans Affairs Department adopt a modified, open-source approach to re-engineering its VistA electronic health-record system.
In one sense, they were telling the VA to go back to the future, since open-source IT development and the VA have had a long history. Still, VistA veterans contacted for this story, including some who have had decades-long relationships with the VistA system, found the recommendations to be a mixed bag.
A work group of the Industry Advisory Council, or IAC, of the not-for-profit American Council for Technology released on Friday its 101-page VistA Modernization Report: Legacy to Leadership after delivering it to the VA earlier last week.
The IAC was tasked by VA leadership last fall to make recommendations on ways to improve the VistA system, which runs at 153 VA hospitals and more than 700 clinics.
Its 42-member work group recommended the VA scrap all of the VistA software code and hand over the task of writing replacement software to a not-for-profit open source foundation. The foundation's job would be to oversee what the IAC work group envisions as a “VistA 2.0 Open Source Core Ecosystem.”
Writing code for the new VistA system will require the collaborative work of in-house VA programmers, giant corporate IT systems contractors, vendors of proprietary commercial-off-the-shelf software systems and small, entrepreneurial open-source software developers, the work group report said.
Its call to upgrade VistA and use an open-source development process to do it was welcome news to many in a growing, open-source VistA community. The work group's chosen method to perform the upgrade, however, rubbed many VistA community members the wrong way, although not all of them for all of the same reasons.
Among the more controversial IAC work group recommendations, at least to several VistA community members, are:
- For the VA to abandon the VistA code base, which community members see as a waste of a public investment.
- The continuation of a trend already under way within the VA of replacing open VistA modules with proprietary, commercial off-the-shelf, or COTS, software products, steps that one VistA developer likened to a “diabetic amputation” of the VistA system.
- Contract out substantial chunks of the new software development work to Beltway IT firms as opposed to in-house development of VistA code at the VA where, historically, clinicians and programmers had worked in close proximity.
- Embrace a more limited set of open-source software licenses that permit privatization of VistA derivative systems.
- Dismiss a more communal form of licensing favored by WorldVistA, an existing, not-for-profit open-source VistA development group, which requires a freer and more open sharing of software code. The license, the WorldVistA folks argue, lowers software purchase prices and ongoing operating costs by eliminating software license fees.
Robert Kolodner, the former head of the Office of the National Coordinator for Health Information Technology at HHS, is a psychiatrist and medical informaticist who worked at the VA for 28 years, most of them in health information technology.
Since October, Kolodner has served as chief health information officer for Open Health Tools, Asheville, N.C. Kolodner chose his words carefully. Skip McGaughey, executive director of Open Health Tools, advised the IAC group on open source and licensing issues.
“I think that this report presents a very interesting alternative for the VA on how to proceed,” Kolodner said. “I look forward to what the VA's reaction is to the report and how it plans to follow through and what parts it plans to proceed to execute.”
Roger Baker, the VA's equivalent of a chief information officer, declined a request to be interviewed for this story.
A request to the ONC for a response on the IAC work group plan was not responded to by deadline. The ONC's role is to consult with and coordinate the health IT activities of all federal healthcare organizations. According to its report, IAC work group members did not consider the effect of their recommendations on the IT programs of the Defense Department or the Indian Health Service, both of which have IT systems derived from that created by the VA.
Other VistA community members were not so reluctant to provide strong opinions about the IAC proposal.
“They have identified some issues that are real issues,” said VistA pioneer Gordon Moreshead, about the need to upgrade deficient VistA elements, including the laboratory record-keeping system Moreshead helped develop. “These are things they should have done a long time ago,” he said.
Moreshead started working at the Salt Lake City VA hospital in 1970, seven years before the arrival at the VA of fellow pioneers Ted O'Neill and Marty Johnson, who, according to VistA history, led the earliest efforts to promote clinical computing at VA hospitals through a program of disbursed software development.
Morehead joined or led teams of developers to envision and then create several key VistA applications, such as labs, computerized physician order entry, or CPOE, and the graphical user interface for the Computerized Patient Record System, the “Windows-like” front-end application of VistA that clinicians see and use to document patient care and pull up patient records.
Moreshead said he's not opposed to overhauling VistA; on the contrary, he's long supported it.
“I was in the first planning meeting with Ted O'Neill,” Moreshead said, “but I proposed architecture changes back in 1995 and 1996 and data standards to support that—I still have the old PowerPoint slides here somewhere—and the VA chose to ignore that, which is one of the reasons why I chose to leave.”
Moreshead said he's concerned with the way the IAC work group has proposed making some of those changes “because they are so skewed” in favor of contracting out new programming and development to outside firms and also because these outside IT experts underestimate the difficulty of the task of re-engineering the VistA system.
For example, Moreshead noted that Ed Meagher, chairman of the IAC work group, and a former VA IT leader, estimated it would take IT contractors a few months to “decompose” the code of the current VistA system, catalogue its core functionality and that of each of its 130 modules and create specifications to guide developers in replicating those functions in updated programming language and architecture in VistA 2.0.
Moreshead points to work on the most recent VistA replacement program, called HealtheVet, which the VistA 2.0 project would supersede, relied heavily on contracted out software development and dragged on for nine years and was never completed.
“I just think it's a terrible mistake,” Moreshead said of the proposed VistA development freeze. “Their five years could take 20, and then what do you have? It's a disaster in the making in terms of the veteran's care.”
[Click here to comment on this article.]