VA Extends Financial System Overhaul by 5 Years—And Adds More than a Billion to the Cost

A core IT modernization effort at the Veterans Affairs Department is billions of dollars over the original budget and the timeline for full deployment has been pushed out to 2030, officials told Congress Thursday.

While the agency’s multibillion-dollar effort to modernize its electronic health records system gets most of the attention, a joint hearing Thursday between the House Veterans Affairs Committee subcommittees on Technology Modernization and Oversight and Investigations focused on the agency’s financial management system.

The Financial Management Business Transformation, or FMBT, program kicked off in 2016 as a shared services project with the Agriculture Department. The agencies worked together to pen a contract to deploy the Momentum financial management system developed by CGI at an initial cost estimate of $887 million and timeline for full deployment by 2025. However, in December 2017, Agriculture pulled out of the deal, leaving VA to continue on its own.

Now, program cost estimates are upward of $2.5 billion and VA revised the schedule to put full deployment out past 2030.

Jon Rychalski, VA chief financial officer and assistant secretary for management, who worked as a contractor on the last failed financial management upgrade—CoreFLS in 2006—told Congress this could be seen as a blessing.

“FMBT arguably had its own rough start,” he told the subcommittees. “What began as a shared services partnership with the U.S. Department of Agriculture abruptly transitioned to a sole-VA effort when USDA pulled out of the deal. This might have been the best thing that happened to this program. We painstakingly built very close relationships with our business partners, both internal and external. I firmly believe we—the collective VA ‘we’—are best positioned to execute this implementation with strong business partners like CGI, Deloitte and many others.”

The financial services overhaul is no small matter for VA, Rychalski said. He offered a scenario to illustrate:

Currently, when a contract is awarded in our contract management system, the award amount is not forwarded to the financial management system for processing. Instead, communication from contracting personnel to finance personnel to manually enter the award amount into the financial management system is required. This redundant keying of information is error-prone and is frequently overlooked, necessitating tedious and time-consuming reconciliation of the contract management and financial management systems. Under FMBT, the financial impact of the contract award action, as well as any subsequent changes to the contract will be automatically recorded and forever linked to the contract award, thus eliminating the need for reconciliations and strengthening the accuracy of financial reporting and increasing the speed and accuracy of payments to contractors.

But the financial system deployment is not happening in a vacuum, he said. The agency is currently in the midst of two other major modernization efforts: the aforementioned EHR modernization and the Defense Medical Logistics Standard Support, or DMLSS, system.

“The accounting system must be tightly linked to the logistics system in order to ensure ordering, delivering and payments are seamless,” he said.

Rychalski pointed to problems with the CoreFLS rollout.

“The core accounting system, logistics system and users could not function together,” he said. “Surgeries had to be canceled because medical supplies and equipment were not available. This has real-life risk.”

Linking to the DMLSS system would be one thing, Rychalski said, but the Defense Department system is in the middle of migrating to a cloud platform. If VA were to launch the FMBT system and connect it to DMLSS now, the team would have to do it all over again in a year.

“If we deploy FMBT quickly, we will need to connect the financial system to the logistics system two separate times with two major training initiatives. Our current schedule has us connecting to the final cloud-based version, LogiCole, one time with one set of training,” he explained. “Faster deployment means more disruption to end-users, technical rework, cost and risk. Slower deployment means it will take longer to get the full benefits of the FMBT system.”

Rychalski said he and the team aren’t happy with a 10-year deployment timeline, and they would like to accelerate that schedule if possible. However, that will depend greatly on the rollout of DMLSS over the next year.

“Through this year, with our deployment to [the National Cemetery Administration] and [Veterans Benefits Administration], with the deployment of the electronic health record and DMLSS, I think we’ll have a better sight, picture of how complicated all of that is, how successful we are with FMBT, how disruptive EHRM and DMLSS is to the sites it’s going to,” he said. “By the end of 2020, we’ll have a much better idea of what is possible.”

Rychalski said he expects to be able to return to the committees around this time next year, once the DMLSS rollout is completed and the FMBT team can begin its implementation efforts. By that time, the team plans to have the new financial system deployed at NCA and be “deep into” implementation at VBA.

Beyond the timeline, the other major shift in the program has been funding. Rychalski said the initial estimate of $887 million through the Agriculture partnership was unrealistic from the start.

“When I came over and saw the cost estimate of $800 million to go systemwide, that just didn’t make sense to me. That was too low,” he said. “What I’ve learned through the years as a consultant and doing CFO functions: Until you get a couple of sites under your belt—some experience—then you know what it’s going to cost. At this point, it’s sort of theoretical.”

Rychalski said the current estimate is around $2.5 billion, though he didn’t have much confidence in that figure.

“I don’t say that because I think the cost is going to go up,” he clarified, but demurred on offering specifics. He noted that changes to the schedule—either moving it up to deploy sooner or pushing it back due to problems with other modernization efforts—will likely lead to that number going up or down.

“The first decision point, for me, is can we successfully deploy this to NCA,” he said. “The next decision point is how smoothly does DMLSS implementation go. … And then, the third question is how disruptive will it be to try to connect FMBT to DMLSS.”