Report: Veteran suicide prevention increasingly looks to use AI, digital capabilities

Veteran suicide prevention programs are increasingly leveraging advanced technologies, such as artificial intelligence, to better identify and support retired servicemembers at risk of self-harm, according to a landscape analysis of nationwide initiatives released on Wednesday.

The report, conducted by the RAND Corporation in partnership with nonprofit Face The Fight, reviewed 156 currently operating initiatives and an additional 226 proposed programs to examine the types of approaches being taken or put forward to decrease the veteran suicide rate. 

Veterans are roughly 1.5 times more likely to die by suicide than those who have not served, with more than 6,000 veteran suicides reported each year for more than the past two decades. The Department of Veterans Affairs, nonprofit organizations and private sector groups have all undertaken various initiatives to lower the veteran suicide rate, including launching programs that seek to predict which veterans are at a higher risk of suicide. 

The vast majority of active programs reviewed in the report were operated by nonprofit organizations, with slightly less than half of these accessed “virtually or via a combination of in-person and virtual access.” Many of the proposed initiatives were identified from the list of applicants for VA’s Mission Daybreak competition — which provides funding to innovative solutions seeking to lower the veteran suicide rate — as well as others who have received grants from Face the Fight. 

The increasing uses of digital solutions, however, coupled with the planned adoption of other emerging capabilities to enhance suicide prevention activities, were identified as factors that could shake up the next generation of crisis intervention initiatives. 

“Advances in technology, including new medical devices and technologies fueled by artificial intelligence, and novel (or newfound evidence for existing) pharmacological approaches also generate questions about the future of veteran suicide prevention,” the report said. 

The analysis found that the most common types of proposed suicide prevention initiatives focus on uses of mobile health applications, real-time monitoring and suicide risk assessment tools. 

AI tools, however, were also identified as taking on a more prominent role in future prevention programs. The report found that 17% of current programs “explicitly stated they use AI in their veteran suicide prevention activities,” while “among proposed programs, this percentage more than doubles” to 37%. 

“These results are unsurprising, given the notable shift from community-based social connection and case management efforts among currently offered programs — in which AI tools would likely be inappropriate or inapplicable — toward the multifunctional digital health platforms, which frequently feature suicide risk identification and real-time monitoring, among proposed programs,” the report said.

Rajeev Ramchand, co-director of the RAND Epstein Family Veterans Policy Research Institute and one of the report’s authors, also told Nextgov/FCW that “what we're seeing is people using a whole bunch of massive amounts of data and then AI to kind of identify who is at risk of suicide.”

Ramchand noted, however, that uses of these emerging capabilities also need to be balanced with a continued focus on next-step interventions.

“Let's say we have this amazing thing that identifies people at risk,” he said. “I think what's missing kind of is the next question, which is, ‘okay, once we identify groups at high risk, what do we do?’ The AI and the machine learning can identify these groups, but that's just the beginning step. So what do we do? And I don't think we have a great evidence base yet as to what we do.”

VA, for its part, has been using an AI-powered predictive tool — the Recovery Engagement and Coordination for Health — Veterans Enhanced Treatment program — that analyzes data from veterans’ electronic health records to identify those in the top 0.1% tier of suicide risk at each medical center. REACH VET coordinators subsequently reach out to the identified veterans to try and provide more targeted care. The department is in the process of rolling out a 2.0 version of the REACH VET model that includes additional risk factors, such as military sexual trauma.