VA Moves AI Beyond Pilots to Improve Veteran Care

The Department of Veterans Affairs is moving artificial intelligence beyond pilot projects to applications that are delivering measurable improvements in veteran care and driving efficiencies across one of the world’s most complex healthcare systems.

“I think two years ago, we were very much still in that R&D phase, building things in house and seeing what works,” said VA Acting Chief AI Officer Kimberly McManus at the Health IT Summit Wednesday. “Today we have a variety of projects and outcomes that could really actually improve veteran outcomes and improve our services. It’s been really exciting to see.”

McManus said VA’s use of AI has quickly advanced beyond early efforts to gain efficiencies in tasks such as extracting information from paper records and transcribing clinical visits. The department is now applying the technology more proactively to some of its most pressing healthcare challenges.

AI-enabled systems can flag veterans at risk of suicide, while other applications have helped reduce opioid overdose mortality by 22% and improve detection of precancerous lesions during colonoscopies by 21%.

Those clinical gains have been accompanied by efforts to streamline administrative operations. VA found that some employees saved up to six hours per week using general-purpose AI tools, while other applications have helped clinicians spend less time on paperwork and more time caring for veterans.

McManus said those gains have been made possible not only by the rapid expansion of commercially available AI capabilities, but also by VA’s growing experience identifying and testing tools that best support the department’s mission.

“How we’re investing is different. Core capabilities, a lot can be done with those that they couldn’t two years ago,” she said. “You can build dashboards, you can write software, ephemeral, on-the-fly-software, in a way that we’ve never seen before. I think that’s really going to change how we approach the role of custom software development in what we buy and how we buy it.”

As VA continues to adopt AI, McManus said the agency is taking a rigorous approach to determine which technologies are best suited for veteran care.

“We spend a lot of time not just on the technology, but having the right change management, the right people at the table from the beginning as well to ensure that with anything we deploy, the benefits dramatically outweigh the risks,” she said. “I think the other thing that is important to note is that there is risk of not deploying technology.”

McManus said VA will soon roll out additional AI tools aimed at improving service delivery by reducing time spent on internal processes. The agency is also investing more in customer success enablement and AI literacy to help its workforce effectively adopt the technology.

“We are going to see more of a productive and preventative healthcare system,” she said. “It’s a really exciting time.”