A leader in policy and legislation and at the helm of Government Curated, Donald Gainsborough provides a unique insight into how emerging technologies can reshape critical areas such as veterans’ mental health and suicide prevention. As we dive into this topic, Gainsborough helps unravel the intricate layers behind the adoption of artificial intelligence (AI) by the Department of Veterans Affairs (VA) in their efforts to tackle this grave concern.
What prompted the Department of Veterans Affairs (VA) to adopt artificial intelligence in their suicide prevention efforts?
The decision to integrate AI into suicide prevention efforts was largely driven by the persistent and troubling statistics regarding veteran suicides. With more than 140,000 veterans having taken their lives since 2001, the VA recognized the potential of AI to identify veterans at high risk of self-harm more effectively. They sought a proactive approach that could analyze vast amounts of data to offer insights beyond traditional methods.
Can you provide an overview of the veteran suicide statistics that have driven these efforts?
The statistics have been quite alarming. In 2022 alone, an estimated 6,407 veterans died by suicide. What’s particularly concerning is that some studies suggest these figures might even be underestimated, citing a rate about 37% higher than what the VA has reported for certain years. These numbers signify a pressing need for enhanced predictive capabilities.
How did the REACH VET program come into existence, and what role did David Shulkin play in its implementation?
The launch of REACH VET, which stands for Recovery Engagement and Coordination for Health-Veteran Enhanced Treatment, can be credited to a shift initiated under David Shulkin. His tenure saw a concentrated effort to move from reactive to proactive suicide prevention measures. Shulkin encouraged the application of predictive models to better manage and intervene in high-risk cases as quickly as possible.
What specific challenges did VA face in suicide prevention before the introduction of the REACH VET program?
Prior to REACH VET, the VA faced significant challenges in prioritizing and targeting its preventive efforts. The strategies in place at the time were largely reactive—they responded to crises as they arose, rather than predicting and intervening early. This lack of a structured, anticipative approach hindered their ability to reduce suicide rates effectively.
Could you explain how the REACH VET algorithm works to identify veterans at high risk of suicide?
The REACH VET algorithm leverages machine learning to sift through electronic health records of veterans, identifying those in the top 0.1% tier of suicide risk. Initially, it considered a broad range of variables — about 381 — but was streamlined to 61 key predictors, thanks to insights from Dr. Ronald Kessler. The algorithm processes these variables monthly to flag high-risk individuals for targeted intervention.
What role did Dr. Ronald Kessler play in refining the REACH VET model?
Dr. Ronald Kessler played a crucial role in refining the model by trimming down the extensive list of variables. He used a statistical technique called Lasso regression to determine which factors most accurately predicted suicide risk, ultimately strengthening the model’s efficacy and focus.
How does the REACH VET program incorporate machine learning to enhance its predictive capabilities?
Machine learning is at the heart of the REACH VET initiative. This technology allows the program to identify patterns within large datasets, something that would be near impossible manually. By analyzing health records, it continually improves its predictive power, learning which variables are most indicative of future risk.
Which specific variables does the REACH VET model consider when assessing suicide risk?
The model considers a host of variables, such as recent suicide attempts, emergency room visits, certain medications, and mental health diagnoses like depression and bipolar disorder. These factors provide a robust framework for predicting suicide risk and shaping intervention strategies.
How does the REACH VET program balance technological innovation with the “human touch” in suicide prevention?
The program maintains a delicate balance by pairing technological advances with personalized human interaction. While AI sets the groundwork by identifying high-risk individuals, it’s the VA’s mental health personnel who engage with veterans, crafting safety plans and providing support, ensuring that human connection remains central.
How are high-risk veterans engaged by VA facilities once they are identified by the REACH VET program?
Once identified, high-risk veterans receive outreach from VA facilities through dedicated REACH VET coordinators. These professionals work directly with the veterans to develop safety plans and offer resources, conducting unscripted conversations designed to address the veterans’ needs and concerns comprehensively.
In what ways has the medical community responded to the results of the REACH VET program?
The broader medical community has largely applauded the REACH VET program. Studies, such as one published in JAMA, have highlighted its success in increasing treatment engagement, decreasing emergency visits, and reducing nonfatal suicide attempts—all testament to its effectiveness.
What other AI and machine learning initiatives are being utilized by VA to prevent veteran suicides?
Beyond REACH VET, the VA explores several AI initiatives, including models predicting overdose risks and utilizing natural language processing to detect suicidal ideation in PTSD notes. These efforts aim to broaden the spectrum of preventative care strategies.
How is the REACH VET program being updated, and what new factors are being included in version 2.0?
Updates to the program include the integration of additional risk factors such as military sexual trauma and intimate partner violence. These enhancements are part of ongoing efforts to ensure the model remains relevant and comprehensively predictive.
What limitations exist within the REACH VET program in addressing veteran suicides?
While effective, REACH VET only targets a fraction of those at risk. The top 0.1% it identifies accounts for just 2% to 3% of all suicides among veterans. There remains a significant gap in addressing those who are outside the scope of this high-risk category.
How does the volume of veterans identified by the REACH VET model vary across different VA facilities, and why?
The number of veterans flagged varies by the size of the population served at each facility. Larger facilities with more veterans naturally identify more high-risk cases. However, discrepancies in support and coordinator caseloads can influence engagement success.
How does VA ensure that other suicide prevention initiatives are integrated with REACH VET efforts?
VA interlaces REACH VET with existing initiatives like the Veterans Crisis Line and partnerships with outside resource organizations. This alignment enables a more comprehensive approach to mental healthcare and suicide prevention.
What are the main challenges veterans face during their transition from military service back to civilian life?
Transitioning veterans often struggle with the abrupt shift from structured military life to the unpredictability of civilian life. This period can be fraught with mental health challenges, exacerbated by feelings of isolation and a lack of seamless support systems.
How does the Transition Assistance Program aim to assist veterans during this transition?
The Transition Assistance Program is designed to ease veterans back into civilian life, offering classes, resources, and guidance on navigating this new chapter. However, veterans report that the program’s effectiveness is limited by its delivery timing and depth of engagement.
How effective have the efforts been to reduce veteran suicides outside of those veterans already engaged with VA services?
Efforts have seen mixed success. While VA-initiated interventions have made strides with engaged veterans, the challenge remains in reaching those not connected to VA resources. Nearly half of the veterans who die by suicide are not active with VA services.
What steps is VA taking to reach veterans who are not connected to its healthcare or benefits services?
Repeated efforts focus on outreach, public awareness campaigns, and collaboration with veteran service organizations to tap into networks beyond traditional VA channels. These initiatives aim to bridge the gap for veterans outside the VA system.
Despite increased spending on suicide prevention, why do you think the veteran suicide rates have remained “essentially unchanged” since 2008?
Several factors contribute, including societal stigmas surrounding mental health, challenges in consistently reaching non-VA-connected veterans, and perhaps the need for novel approaches that adapt to evolving veteran demographics and conditions.
What future plans does the VA have to enhance its approach to veteran suicide prevention?
The VA is exploring a fresh suite of preventive strategies, already under discussion at the congressional level. There’s a continued push for innovation in AI applications, collaboration with mental health communities, and improving transition support mechanisms to better support veterans before crises develop.
