WASHINGTON, DC—Concerned about several high-profile, high-cost IT project failures over the last several years, legislators called VA officials to Capital Hill last month to provide reassurance that the agency has learned from its past mistakes. While VA IT leaders could not guarantee that no such failures will occur in the future, they did outline steps VA has taken to ensure that future failures will not be nearly as expensive.
“VA has stumbled over the years on its path toward the goal of an electronic VA,” declared Sen Daniel Akaka, D-HI, chairman of the House VA Committee last month. “Most recently, we had a financial and logistics system fail [that was] known as CoreFLS. To make matters worse, the contractor responsible was paid a bonus. Software systems processing GI Bill claims suffered many false starts. Last summer, VA halted 45 projects that were dramatically over budget and overdue, including an outpatient scheduling system that was three years overdue.”
VA needs a clear vision of IT reform with day-to-day management that is working towards that vision, Akaka declared.
Recurring Themes of Mismanagement
Problems with VA IT projects garnered national attention in 2004, when VA pulled the plug on the creation of a next-generation financial system called CoreFLS that was being tested at the Bay Pines VAMC. After two years of testing and $472 million, it was determined that the system was a failure and that it suffered from severe mismanagement. The episode led to the resignation of then-Under Secretary for Health Dr Robert Roswell.
To compound matters, the project that replaced CoreFLS—the Financial and Logistics Integrated Technology Enterprise (FLITE)—was found by VA Office of Inspector General investigators to be suffering from some of the same management problems that plagued the CoreFLS project. OIG investigators discovered a lack of guidance, staffing shortages, and management that did not take well-timed actions to ensure the cost, scheduling, and performance goals stayed on track.
“These reoccurring themes have repeatedly hindered VA’s Office of Information and Technology from being able to adequately develop their system,” Belinda Finn, assistant VA IG for audits and evaluations told the committee. “We issued three reports on FLITE [and] concluded that program managers were repeating problems from the CoreFLS project.”
New OMB guidance on financial systems IT projects, issued in June 2010, had a major impact on the FLITE program. Large-scale federal financial system modernization efforts have historically led to complex project management requirements. Due to their length, by the time the projects are finished they are frequently technologically obsolete. Consequently, OMB has directed all federal chief financial officers to immediately halt new procurements for such projects until new project plans can be developed by the agencies and approved by OMB.
Because of this guidance, Roger Baker, VA’s assistant secretary for IT, announced in July the termination of the integrated financial accounting system and data warehouse portions of the FLITE project.
Baker pointed to this new willingness by VA to terminate projects that are proving untenable as proof that VA will never have a failure like CoreFLS again. “I think the biggest lesson we took from the failure of the replacement scheduling application is that we have to make certain the hard decisions are made,” Baker explained. “You’ve seen a series of hard decisions made at VA relative to other projects. Stopping 45 projects in July of last year was frankly a hard decision—facing that those projects were not deliverable.”
Those 45 ongoing and failing IT projects terminated in July 2009 generated $54 million in cost avoidance in 2010, Baker said. And VA substantially decreased the risk of failure in 33 other projects by replanning and reforming them.
“Reforming a few of them was not viewed positively, but we recognized that they were not going to deliver if we did not change them to an incremental delivery,” he declared.
As for the decision to cease work on major portions of FLITE, Baker said, “In the end we have to determine, ‘Can we be successful? And if we believe we can’t, if we believe it’s an overreach, we have to not do the program.”
New policy by VA does not allow a project to move forward that does not have a “customer-facing deliverable” within the next six months, Baker said. “What that means is that these projects are not going to go a long time, like the replacement scheduling project did.”
VA was forced to cease work on a multi-year replacement scheduling application development program in 2009 because of what OIG investigators found to be ineffective planning and oversight. “Replacement scheduling went years without delivering anything before they figured out they couldn’t deliver anything,” Baker explained.
By requiring projects to have a producible goal within six months, VA is limiting how long a project can go before being recognized as failing—a technique Baker referred to as “fail fast.”
“If it’s going to fail, we want to figure it out quickly and stop spending money on it. That has generated a lot of facing up to those hard decisions within the organization,” he admitted.