AI alerts improve suicide prevention in clinics

Over the past decade, there’s been growing concern about the staggering rates of suicide. Now, a remarkable study from Vanderbilt University Medical Center (VUMC) offers a ray of hope. The study illustrates how artificial intelligence (AI) alerts can aid doctors in identifying patients at a higher risk for suicide.

Led by Dr. Colin Walsh, the research team tested the Vanderbilt Suicide Attempt and Ideation Likelihood (VSAIL) model. This AI-driven system aimed to prompt suicide risk screening in three VUMC neurology clinics.

The findings, published in the journal JAMA Network Open, demonstrate that interruptive alerts – which actively notify doctors during their workflow – were far more effective than passive notifications embedded in electronic charts.

AI-driven alerts for suicide risks

The study compared two types of AI-driven alerts. Interruptive alerts actively interrupted the doctor’s workflow by appearing as pop-up notifications during a patient consultation.

This required the doctor to acknowledge and address the alert immediately, thus ensuring that they took action right away.

Passive alerts provided the same risk information but did so in a less direct way. The information was displayed within the patient’s electronic medical chart, where it could be seen but did not actively prompt the doctor to take immediate action.

This approach avoided interrupting the workflow but relied on the doctor noticing and acting on the information independently.

The study found interruptive alerts to be more effective at prompting doctors to conduct suicide risk assessments.

Interruptive alerts reduced suicide risk

The results were striking. The team found that interruptive alerts led to suicide risk assessments in 42% of cases, while passive alerts resulted in assessments in only 4% of cases.

“Most people who die by suicide have seen a health care provider in the year before their death, often for reasons unrelated to mental health. But universal screening isn’t practical in every setting. We developed VSAIL to help identify high-risk patients and prompt focused screening conversations,” said Dr. Walsh.

Suicide: A growing public health crisis

Suicide rates in the United States have been increasing steadily, with 14.2 deaths per 100,000 people annually. Suicide is now the 11th highest cause of death nationwide.

The researchers noted that 77% of individuals who die by suicide have seen a primary care provider within the year before their death. These figures highlight the critical need for better ways to identify and support individuals at risk.

The VSAIL model offers a significant advance in addressing this challenge. It uses data from routine electronic health records to assess a patient’s 30-day risk of attempting suicide.

In prior testing, VSAIL demonstrated its effectiveness by identifying high-risk individuals, with one in 23 flagged patients later reporting suicidal thoughts. This capability positions VSAIL as a powerful tool for targeted suicide prevention efforts.

Identifying high-risk patients

The new study involved 7,732 patient visits over six months, which triggered 596 suicide alerts.

Researchers focused on neurology clinics, as certain neurological conditions carry heightened suicide risks. Of the flagged visits, only about 8% prompted alerts, highlighting the model’s efficiency in busy clinical environments.

During the 30-day follow-up, no flagged patients reported suicidal ideation or attempts. However, the team noted potential downsides, such as “alert fatigue,” where frequent notifications could overwhelm clinicians. Future studies will explore this balance.

“Health care systems need to balance the effectiveness of interruptive alerts against their potential downsides,” Walsh noted.

“But these results suggest that automated risk detection combined with well-designed alerts could help us identify more patients who need suicide prevention services.”

Broader implications of suicide alerts

The study’s success suggests that similar systems could benefit other medical settings. By selectively flagging high-risk patients, AI models like VSAIL offer a feasible and impactful approach to suicide prevention.

“The automated system flagged only about 8% of all patient visits for screening. This selective approach makes it more feasible for busy clinics to implement suicide prevention efforts,” Dr. Walsh concluded.

The research was conducted by a multidisciplinary team from VUMC, including Dr. Michael Ripperger, Dr. Laurie Novak, and co-senior authors Dr. William Stead and Dr. Kevin Johnson.

The study paves the way for innovative, AI-driven interventions that could save lives and redefine suicide prevention in healthcare.

The research is published in the journal JAMA Network Open.

—–

Like what you read? Subscribe to our newsletter for engaging articles, exclusive content, and the latest updates. 

Check us out on EarthSnap, a free app brought to you by Eric Ralls and Earth.com.

—–