I was part of a small group of programmers (called Hardhats) recruited to the VA in the late 1970’s to work on what would eventually become the VistA Electronic Health Record system. Ted O’Neill, who had supported the funding of the development of ANS MUMPS from NIH and later National Bureau of Standards (now NIST), moved to the VA to develop an open, public domain version of a modular, decentralized hospital information system that was dedicated towards improving the clinical care in the VA.
This was in an era dominated by mainframe computers, managed by centralized data processing staffs doing largely batch processing of punched cards. The notion of a network of interactive terminals connected to decentralized minicomputers was a radical notion at the time, and was threatening to the centralized data processing department.
This lead to a fierce bureaucratic battle between the decentralists and the centralists. Ted O’Neil and Marty Johnson hired MUMPS programmers under local hospital management, both to insure that they worked closely with the actual end users, and to shield them from the conflicts that raged in Washington. Eventually, Ted O’Neill was fired, several of the hardhats were fired, and central office tried to shut down the MUMPS effort. I was demoted, and $500,000 worth computers were locked up in my hospital basement, unused. The central data processing department told upper VA management that minicomputers could not possibly be used for large scale computing, and that only a centrally managed mainframe approach could provide the necessary functionality.
The hardhats continued to develop the software, cooperating on a peer-to-peer basis, and working closely with hundreds of doctors, nurses, and other clinical personnel. By 1981, we had developed a toolkit (the File Manager, Kernel) that supported a core system that could handle packages for ADT (Admissions, Discharges, and Transfers), Pharmacy, Scheduling, and Laboratory.
In 1981, VA Chief Medical Director Donald Custis visited the Washington VA medical center to see our software in operation. He was surprised to find a working system, enthusiastically used by clinical staff, based on very economical minicomputers. He quipped, “It looks like we have an underground railroad here.” I grabbed the name, and started passing out 500 VA Underground Railroad business cards.
In 1982, I organized the first Underground Railroad banquet in Washington, DC, and presented then-Deputy VA Administrator Chuck Hagel with an “Unlimited Free Passage on the Underground Railroad” certificate. I also started handing out certificates for “Outstanding Engineering Achievement” to programmers for their contributions to VistA, and special VIP membership cards, with a 1982-era Motorola CPU chip laminated to the engine of the logo.
I am planning the next banquet October 24, 2013 in conjunction with the VistA Expo meeting in Seattle. I will be delivering a “State of the Underground Railroad” address, discussing how many of the original issues are still around, 31 years later.
For example, I had noted that in a bureaucracy, everyone wants things centralized below them and decentralized above them. Given the technology of the day, we focused on the hospital as the “anchor point.” Today, however, this has moved up to Capitol Hill. Both Senate and House committees have discussed what language to use in EHR systems. The $1b disastrous Integrated Electronic Health Record effort is an effort in mega-centralization. DoD continues it’s Humpty Dumpty systems development approach, breaking systems into pieces and then trying to integrate them back together again, even after a 40 year track record of failure.
VistA’s approach to a patient- and provider- centric model has repeatedly proven it’s merit. Our approach of involving thousands of clinical users – not just a few IT “experts” – has also proven itself. Open source software, agile development, use of online fora, metadata-driven architectures, and email-based messaging are all innovations of VistA that are more current than ever.
VistA was much more than just a collection of programs. It was a community of users, a framework for collaborative development, and a toolset for “meta” level programming that is rarely understood by outsiders who stare at the source code. Just as one cannot understand Wikipedia and the Wikipedian community by staring at the source code driving the underlying wiki, we cannot understand VistA simply by looking at the source code.
I hope that the Underground Railroad Banquet can help communicate some of these broader implications of the VistA framework, as well as look forward to the next generation of VistA software.