As the COVID-19 pandemic continues into autumn, the Department of Veterans Affairs (VA) has continued its effort to support veterans in crisis by partnering with customer experience management provider Medallia to launch the Veterans Signals program.

The Veterans Signals program collects feedback from veterans via survey, and uses AI capabilities to  analyze the tone and language of the veterans— allowing VA to respond to those that may be at-risk. According to the Centers for Disease Control and Prevention (CDC), the COVID-19 pandemic been associated with an increase in mental health challenges and suicidal ideation.

The Veterans Signals program has so far routed 2,800 crisis alerts to VA offices and enabled the agency to provide early intervention services to more than 1,400 veterans within minutes.

To discuss the Veterans Signal program, MeriTalk caught up with Lee Becker, who is Solutions Principal at Medallia and former chief of staff for the Veterans Experience Office at VA. He highlighted that the solution’s AI and machine learning capabilities allow for the distillation of massive amounts of unstructured data.

Becker said that Medallia’s technology aims to “take massive, massive amounts of unstructured data, and be able to very quickly, instantaneously, really distill what the actual insights are, and use that to drive action,” Becker said.

“So from the standpoint around suicide prevention as a use case, we would provide this capability to veterans to provide input on their experience around their care, benefits and services. Veterans who have shared that they were thinking about self-harm, or they’re homeless, they’re in crisis, they would use any type of various key words. And the technology that Medallia provides – which offers better signals for VA – is able to distill that instantaneously and identify those veterans that are in crisis in a very secure environment.”

The COVID-19 pandemic has led to VA using topic sets around “COVID” to sharpen response. The increase in telehealth services across the healthcare industry also has complemented VA’s efforts.

When asked about any potential AI bias that may be learned through the system, Becker said that has always been a consideration, and awareness of possible bias is important for maintaining trust between VA and veterans.

“The way we look at this data, we really are looking at it from a segmented perspective which is so important. As a best practice we encourage looking at race, ethnicity, age, gender, and sexual orientation, looking at that whole gamut. I think that is really key,” Becker said. “We need to understand it from all those different aspects and make sure that we check the system’s biases to make sure that we’re not being led down the wrong path.”

Becker said VA is aware that trust among women veterans is much lower than trust of the VA among male veterans. That’s something that VA wants to continuously be aware of in order to provide the best care, Becker said.

Going forward, Becker said VA is looking to expand the program to be inclusive of veterans’ family members and caregivers as they are important members of the veteran ecosystem, and are able to capture signals from them that will be helpful for care. He also said VA is doubling down on the employee aspects of the program with veteran’s identifying employees in their feedback.

“Of course, it’s about technology, right? We need to use this technology, but really, we need to consider the people aspect, the process aspect,” Becker said. “And what I implore leaders to do is really be able to orient their ability to listen to their employees to help them, and help them enable the technology transformation they’re trying to do.”

Read More About
More Topics
Jordan Smith
Jordan Smith
Jordan Smith is a MeriTalk Senior Technology Reporter covering the intersection of government and technology.