Members of the not-for-profit organization WorldVistA gathered for three days last week at the 14th VistA community meeting near Washington.
Rolling through the 578-acre campus past the many buildings of the National Institute of Standards and Technology in Gaithersburg, Md., one wonders, given all the resources devoted to standardization, how any problems with standardization of healthcare information technology could remain unsolved.
Yet in fairness to the institute, which is home to two Nobel Laureates, there is a lot on its test bench, even though healthcare IT has long been a big part of the NIST portfolio.
And as such, it was something of a homecoming for WorldVistA, whose members promote use of the Veterans Health Information Systems and Technology Architecture, or VistA, clinical record system for use in physician offices and hospitals outside the Veterans Health Administration.
The precursor to NIST, the National Bureau of Standards was founded in 1901. Three years later, the bureau faced an early challenge when a two-day fire destroyed an estimated 2,500 buildings in Baltimore.
Firefighters from New York City, Philadelphia and Washington called in for backup watched in frustration as 80 square blocks of the city burned because their hose connections didn't match up with Baltimore hydrants. Afterward, the bureau examined 600 couplings from fire departments across the country to come up with standards for threading and a universal connector. The federal agency has been dealing with all manner of manufacturing standards and quality improvement programs ever since.
The bureau would play a significant role in laying the foundation for today's healthcare IT infrastructure, since many of the major healthcare IT systems are based on the Massachusetts General Hospital Utility Multi-Programming System, or MUMPS, database and programming language.
Thirty years ago, Joseph Ted O'Neill and Martin Johnson were with the bureau, working as part of a public-private effort to get a MUMPS programming language standard approved by the American National Standards Institute. Later in 1977, O'Neill and Johnson came to work as part of the computer-assisted system sttaff within the VA's Department of Medicine and Surgery.
There, they had the vision that MUMPS could be brought into the VA hospitals nationwide, wrote George Timson, a pioneering former VA programmer and a historian of the VAs clinical IT effort.
Settling on MUMPS as a common language and database opened the door for Timson and a group of fellow insurgent programmers to incorporate MUMPS in the decentralized development of a clinical computing system across the VA.
In 1978, Timson began work on FileMan, a package of MUMPS programming utilities which, though much expanded, is still in use today within VistA. Timson is still writing code, including a significant May 2006 update of FileMan released as open source by Medsphere Systems Corp., which sells a modified version of VistA to private- and public-sector hospitals and nursing homes.
Members of the loosely knit VistA community include programmers, IT consultants, pharmacists and more than a few geek docs who combine their clinical knowledge with programming expertise. In 2002, they incorporated WorldVistA as an only slightly more structured not-for-profit to provide a legal entity for their future work.
Next month, WorldVistA plans to submit to the federally funded Certification Commission for Healthcare Information Technology its VistA Office Electronic Health Record for testing and, if all goes well, certification. The submission will be a milestone in the life of the project begun in 2004 by the VA and the CMS. The goal of the project is to make EHRs more affordable to physicians in solo practice and small groups, where EHR adoption remains persistently low and where the real-dollar incomes of many primary-care physicians has shrunk during the decade. VistA was a natural choice by the CMS for two reasons. It is in use at more than 800 outpatient facilities across the VA, and the VistA software is in the public domain, thus it is available from the VA without charge under the Freedom of Information Act. The CMS hired the Iowa Foundation for Medical Care, the state quality-improvement organization, to oversee the project, and the foundation eventually hired WorldVistA to help.
WorldVistA members, with funding from the Honolulu-based Pacific Telehealth and Technology Hui, a partnership between the VA and the Defense Department, had developed a branch of VistA that runs on open-source versions of MUMPS and the Linux operating system. VistA Office EHR, or VOE, is based on this open-source VistA software. According to Mark Leavitt, chairman of the certification commission, none of the 37 EHR systems that have received CCHIT certification thus far is open-source.
With all this history and success behind them, WorldVistA members face a crucial turning point in addition to certification testing.
On Jan. 31, the $4 million CMS contract with Iowa Foundation for Medical Care, and thus the foundation's contract with WorldVistA, runs out. The biggest challenge, according to WorldVistA member Cameron Schlehuber, will be to keep up with the VA patch stream, an electronic outpouring of upgrades and bug fixes to VistA that VA programmers produce at the rate of 500 to 600 a year.
One of the discussions that is under way as part of the CMS contract is we have to be able to stay ahead of that VistA tsunami, said Schlehuber, who holds a doctorate in neuroscience and helped develop the VA's computer system from 1978 until he retired last September. We're hoping that the VA recognizes their obligation when they partnered with CMS, that they're responsible for maintaining it.
It should not be a high-cost process, Schlehuber said, noting that it should take 15 to 20 hours of programming time a month to keep up with the stream. The risk is if VOE falls behind the patch stream, the two programs will fork, or go on separate development paths, a potentially costly problem that VistA developers already know all too well.
Like VOE, the clinical IT systems in use today at the Indian Health Service and the Defense Department were based on copies of the VA's clinical IT system. But the IHS system and the $1 billion development at Defense were allowed to fork so badly that records could not be readily transmitted between them and the VA. Since then, taxpayers have paid many millions of dollars to try to make the VA and Defense systems interoperable. Thus, the forking of the Defense system ranks as one of the bigger mistakes in healthcare IT history.
The other challenge, Schlehuber said, is for the VA to create an orderly process for incorporating improvements to VistA developed outside the VA. Several modifications to VistA done in the development of VOE have been presented to the VA, but have yet to be incorporated and distributed via the patch stream throughout the system. The Instituto Mexicano del Seguro Social, one of Mexico's national health services, has installed a Spanish language version of VistA at 56 of its hospitals and has built what programmers there think are improved versions of several VistA software modules. Medsphere has several hospital installations under way in the U.S. using its version of VistA, yet Timson's upgrades to FileMan, created when he was employed by Medsphere, were not readily accepted by the VA. Other VistA vendors, aiming to leverage open-source VOE and target the 200,000-physician small-office market, could boost development of the software dramatically and produce a stream of improvements to the software just as steady as that now emanating from the VA.
There are going to be a number of users of VOE that far exceeds the number of users at VA, Schlehuber said. At some point I can see that is going to develop stresses in VA.
WorldVistA member Brian Lord is a former VA programmer who has worked on the development of VOE. Lord is president of Sequence Managers Software, which is selling a subscription EMR service based on VistA that is distributed to clients using the application service provider, or ASP, model.
WorldVistA plans to release VOE under the Free Software Foundation's General Public License, and that's as it should be, according to Lord.
We are dedicated to open source because we believe there is a business model there, he said. We've got a PACS (picture archiving and communication system) attached to our VistA, a practice-management system that's attached, and it's all open-source.
This product must be released under GPL, Lord said. It has to be community property. They want to ensure that anybody that wants to start working with VistA has to contribute to the community. Unless they are invested in the community, they are not going to be able to make all of this work. The fact is, the intellectual capital that is required to make VistA work and make it grow, no company can match that.
The underlying goal for all of this is to change the way healthcare is practiced and that is to create a system where records can be exchanged, and VistA is a foundation that can make that happen, he said.