The Department of Veterans Affairs is adding additional risk factors to its AI-powered tool for identifying veterans at high risk for suicide to better account for women’s experiences.
The effort comes after a report released in March by the nonprofit Disabled American Veterans warned that the department’s suicide prevention tool – the Recovery Engagement and Coordination for Health – Veterans Enhanced Treatment program – uses retired male military personnel as its foundation used.
The organization noted that the REACH VET model does not take into account military sexual trauma, or MST. VA data shows that one in three women and one in fifty men have confided to their clinical provider that they have experienced MST. The report recommended that VA revise its algorithm to include risk factors for MST and intimate partner violence.
A subsequent study conducted by The Fuller Project, in collaboration with Military Times and Military.com, also found that the program’s algorithm considered being a white male more of an indicator of potential self-harm than other factors that affect women completely or largely.
Naomi Mathis, assistant national legislative director of Disabled American Veterans, shared Next Government/FCW that as VA adopts modern capabilities like REACH VET, “you would think that these tools would improve the VA system or make it work better for the modern service or the modern veteran.”
Fully launched in 2017, REACH VET uses a predictive model to analyze data from veterans’ electronic health records to identify those in the top 0.1% of suicide risk. REACH VET has identified approximately 6,700 veterans each month for additional health care assistance.
The program currently uses 61 indicators in six different categories to identify veterans at risk of self-harm, including based on “demographics, diagnoses, medications, uses, and interaction terms, such as the interaction between marital status and gender.”
A VA official said Next Government/FCW that the department is “in the process of updating REACH VET’s predictive algorithm to consider additional variables specific to female veterans.”
The new risk factors being considered include intimate partner violence and MST, as well as medical conditions that affect women, such as pregnancy, fibroids, endometriosis and ovarian cysts.
“As we update the model, it will be evaluated for performance and bias before deployment,” the spokesperson said, adding that the goal is to launch the new algorithm in early 2025.
VA has been working in recent years on new approaches – such as REACH VET – to support veterans in need of mental health care, even as the number of retired military members who have died by suicide remains high.
According to VA’s 2023 National Veteran Suicide Prevention Annual Report — which was based on 2021 data — the suicide rate among veterans increased 11.6% from 2020. The same study found that the age-adjusted suicide rate increased 24.1 % increased among female veterans from 2020 to 2021, compared to a 6.3% increase among male veterans during the same period.
Mathis said it was a positive step that VA finally moved to add MST as a risk variable to its upgraded REACH VET algorithm, but questioned why it didn’t include it in its original model, especially since it affects both men and women .
“You’re under-reporting MST and you’re saying, ‘MST is not statistically significant,’” she added. “Well, then you miss all those people who feel comfortable reporting it.”
The VA representative pushed back on the claim that the current REACH VET tool prioritizes men, saying it only prioritizes individuals at the highest risk of self-harm. However, they noted that “sex is included as a variable in the model and that being male does indeed have a positive value.”
The Department has already submitted the REACH VET predictive model and use case for a “safe AI assessment,” as outlined in President Joe Biden’s October 2023 Executive Order on the Safe and Trustworthy Use of AI.
“We have tested the current REACH VET algorithm and are also currently testing all candidate models for the new REACH VET algorithm to ensure they fairly assess suicide risk for key demographic populations,” the spokesperson said.
Concerns that specific suicide risk factors are not included in REACH VET’s predictive algorithm have also attracted the attention of some top lawmakers.
Senate Veterans Affairs Committee Chairman Jon Tester, D-Mont., introduced legislation on Sept. 25 to improve mental health care for veterans, with the bill also seeking to address some of the concerns that have been raised around the factors included in the REACH VET algorithm.
The legislation included a provision that would require VA to modify REACH VET to include “risk factors weighted for women,” including MST and intimate partner violence.
Tester said in a statement to Next Government/FCW that VA must add the variables to its model to ensure that all veterans receive the mental health care they need.
“Women veterans are the fastest growing group of veterans, and my legislation will ensure that VA considers the experiences of women veterans, especially survivors of military sexual trauma and intimate partner violence, when treating veterans’ mental health to ensure ensuring no veteran falls through the cracks,” he added.
