Part one of a multipart series (Access part two, part three and part four):
The Veterans Affairs Department should scrap virtually all of the software code in its vaunted VistA clinical health information technology system, an IT industry advisory group recommended last week
Even so, at least part of the legacy of what could fairly be described as the consummate healthcare legacy IT system should survive, according to the group.
Instead of dumping the VistA system entirely, the group recommended that the VA “harvest” VistA by taking the system apart, defining the functions that developers and clinicians have spent more than 30 years trying to perfect, and then using what they learn to create specifications from which programmers can rewrite the code to replicate VistA using a more modern programming language and architecture.
The 42-member work group of the Industry Advisory Council, or IAC, of the American Council for Technology released its VistA Modernization Report: Legacy to Leadership
, advising the VA on how to meet the future IT needs of the nation's largest integrated healthcare delivery system with its 153 hospitals and 768 outpatient clinics.
Industry giants Computer Sciences Corp., Deloitte, Hewlett-Packard Co., IBM Corp., Intel Corp., Lockheed Martin Corp., Microsoft Corp., Northrop Grumman Corp. and Vangent along with lesser-known firms provided representatives to the work group.
Replacing VistA could cost the VA about $5 billion to $7 billion, according to Ed Meagher, the chairman of the IAC work group that prepared the report. The report itself did not include a cost estimate.
Meagher is the director of strategy for health affairs in the global health sector of multinational IT consultant SRA International. Meagher was hired by SRA in 2008 to focus on veterans' and military health issues, according to a news release
at the time. He is a former deputy chief information officer of the Interior Department and deputy CIO at the Veterans Affairs Department, having served at the VA for seven years, leaving in 2006.
Most likely, contracts and assignments for the massive, multibillion-dollar software development project at the VA would be divvied up among giant IT contractors, veteran-owned small businesses, makers of commercial off-the-shelf software products, open-source software developers and VA-employed programmers and clinical personnel, according to Meagher.
“All the approaches up until this one have been evolutionary,” Meagher said. Those prior efforts to improve VistA, he said, were characterized by plans to “go slow, change just a little and improve just a little every day or every month, and that's the safest way to get there.”
Those attempts all failed, Meagher said, including a decadelong VistA replacement effort called HealtheVet, which cost almost $600 million from fiscal 2001 through fiscal 2007 and had its completion date pushed back from 2012 to 2018, according to a 2008 report by the Government Accountability Office. The VA subsequently has estimated an $11 billion price tag for completion of HealtheVet.
“We've tried to modernize this puppy,” Meagher said. “We've tried code conversion. The VA has tried just about every approach to this. It's just too hard of an approach. HealtheVet was announced 10 years ago, and it's just nowhere. The only thing they got up and running was My HealtheVet,” the VA's popular patient portal.
In addition, Meagher said, VistA is “incredibly expensive” to maintain, costing the VA “roughly $500 million a year just to keep it alive.”
The underlying database and programming language integral to VistA is Cache, by InterSystems Corp., an updated and proprietary version of the late 1960s-vintage Massachusetts General Hospital Utility Multi-Programming System, commonly referred to as MUMPS and sometimes simply, “M.”
MUMPS “had its day,” Meagher said. “It does some things very well, but it's just not something you'll want to go forward with.”