If the Veterans Affairs Department chooses to pursue an open-source software development approach to upgrade its VistA clinical information technology system, a key decision must be made: What database software will be used in the redeveloped VistA system?
Where does VistA, Veterans Affairs go from here?
To the VistA outsider, the question—if not the choice—seems simple enough. To members of the burgeoning community of VistA insiders, however, the battle for the answers is shaping up to be a holy war for the digital soul of the software. And, as sometimes happens in soulful battles, there is a lot of money at stake.
Brian Lord is a former VA programmer who is now CEO of Sequence Managers Software, a Durham, N.C., developer of open-source VistA systems.
In reading the report released last week by the Industry Advisory Council, or IAC, of the American Council for Technology advising the VA on a way forward for VistA, Lord said he was “thrilled to see that they even went so far as to say, if you're going to make it open-source, you have to create a community, you have to have a bill of rights to what this community is going to be. I've never seen open-source characterized so well in any political document. That's unheard of.”
Lord says he also favors the IAC recommendation to perform a thorough, module-by-module analysis of the functionality of today's VistA system.
That analysis is “an absolute necessity to move VistA forward,” Lord said. “Too many things are not documented properly. The documents are old. They've not been brought up-to-date.”
Lord breaks with the IAC, however, when it dismisses any future use of the Massachusetts General Hospital Utility Multi-Programming System, commonly known as MUMPS, the database and programming language used in VistA, as outdated technology. MUMPS is among the oldest of the programming languages, but Lord said several vendors have kept it up-to-date.
“I hear a lot of the anti-MUMPS folks, and nine times out of 10, it's just because they feel they cannot hire MUMPS programmers,” Lord said. MUMPS programmers are available, he said. They don't come cheap, but, “it costs money to hire any kind of a programmer.”
“When you start to get into the prejudices, that's when you make bad decisions,” Lord said, dismissing the IAC insistence on using different, newer programming languages and database systems for VistA 2.0 as “just trying to go with what's popular in the market right now.”
IAC work group Chairman Ed Meagher, a former top IT officer for the VA, estimates it will take five years to complete the VistA re-engineering program as outlined in the group's report, and cost $5 billion to $6 billion. The VA has in the works an IT contracting authorization called T4, which is short for Transformation Twenty-One Total Technology, that carries an upper limit of $14 billion. Meagher and the work group want to scrap all of the VistA code and rewrite using something other than MUMPS. But Lord is adamant in saying that would be a mistake.
Meagher “thinks he's going to solve a lot of problems using newer technology,” Lord said, but to get the same performance under a Structured Query Language, or SQL database, “it's going to take three times the hardware that they now have because these new systems do not scale the way MUMPS does. There is no argument that the biggest databases in the world run on GT.M, and that's MUMPS.”
GT.M is an open-source version of MUMPS offered by Fidelity National Information Services.
Send us a letter
Have an opinion about this story? Click here to submit a Letter to the Editor, and we may publish it in print.