VA's early reviews could keep HealtheVet out of sick bay
- By Mary Mosquera
- Apr 22, 2005
'The more problems you find up front, before you even lay pen to paper or write a line of code, that's good.'
'VA CIO Robert McFarland
Henrik G. de Gyor
The Veterans Affairs Department doesn't at the moment have an IT problem with the HealtheVet patient and clinical care system, grumbling in Congress notwithstanding. And CIO Robert McFarland wants to make sure it stays that way.
McFarland's plan is to locate and repair any potential problems before the first line of code is written.
HealtheVet is an online system designed to replace VA's legacy Veterans Health Information Systems and Technology Architecture program, or VistA, an electronic medical-records and clinical-care system used throughout VA's 1,000 plus medical facilities.
To identify potential problems with the system, VA hired Carnegie Mellon University's Software Engineering Institute last year to conduct a technical review before proceeding to system development. SEI, which finalized the report in February, found many potential problems'to the point the project may not be viable as originally planned. So VA went back to the drawing board.
'We did it right this time. We decided to take our preliminary project plan and have experts look at it to help us find the holes and identify the risks so we could come away with a better project plan that we could execute,' McFarland said.
He is optimistic that the new system will proceed because he vetted its design plans through an independent technical review. Despite learning that the system had numerous and significant weaknesses, he painted the technical review as a diagnostic tool to prevent future problems.
'The more problems you find up front, before you even lay pen to paper or write a line of code, that's good,' he said.
The results of the review are 'a more robust systems and engineering process and deciding the target infrastructure,' McFarland said. VA has since begun to modify its development plans and given its revisions to SEI to evaluate.
The SEI study addressed the design, organizational structure and testing of HealtheVet. Among its findings, it said the agency must:
- Create an enterprise architecture and documented road map
- Strengthen the program management office staff and operations procedures to handle a large systems integration project
- Develop technical, operational and management baselines
- Stick to software lifecycle processes and procedures.
The new system will be less expensive to maintain than VistA, have more ability to interface with other applications and be accessible by all veterans.
'This review was not due to a problem that had materialized yet,' said Steven Palmquist, chief engineer for civil and defense agencies, acquisition support board.
Disclosure of the HealtheVet study comes after VA shelved its $372 million Core Financial and Logistics System, or CoreFLS, last year because the pilot deployment did not work properly at VA's Bay Pines, Fla., hospital.Trust but validate
As a result of lessons learned from the CoreFLS failures, McFarland sought independent validation and verification of the HealtheVet plans before beginning development, he said.
In fact, McFarland said, VA will conduct IV&V on most major IT projects very early in the planning.
CoreFLS and now SEI's evaluation on HealtheVet attracted the attention of lawmakers.
The House Veterans' Affairs Committee had already recommended withholding the $311 million request for HealtheVet in the fiscal 2006 budget until questions about the system's management and operability are resolved.
It recommended an overall reduction of $400 million for VA's IT programs, which includes the $311 million for HealtheVet, until VA completes a systems review.
McFarland conferred with House and Senate staff members to quell the rising criticism and received positive feedback on VA's actions.
SEI's Palmquist said a technical review early in the program could save significant time and money later.
VA needs to get started replacing the legacy health care system. VistA is written in M, an old computer language for medical programming. The cost to keep the 20-year-old system afloat is very high, and at some point M programmers will disappear, McFarland said. HealtheVet will move VistA functionality to Java.
VistA modernization is not designed to change the functionality.
'I believe if we change the functionality of VistA in the rewrite that docs will come after us with scalpels,' McFarland said.
Mary Mosquera is a reporter for Federal Computer Week.