Every 43 seconds, someone loses a person who matters —  a friend, a family member, a co-worker — to suicide. Victims of suicide in the U.S. often see a primary care or mental health service provider within the year before their death. 

Are providers reaching people too late or not at the right moment? Statistics make it clear that those at risk for suicide urgently need health care practitioners to think about risk and prevention in new ways, beyond the constraints of a typical health care setting, and especially at times when people are most vulnerable. 

A new study from researchers at the Center for Healthy Minds, published in the journal JAMA Network Open, sheds light on some new ways to understand suicide risk in adults. 

Lead researchers Dr. Ross Jacobucci, CHM research assistant professor, and Dr. Brooke Ammerman, CHM affiliated faculty and assistant professor in the University of Wisconsin–Madison Department of Psychology, found that nighttime smartphone habits can indicate if someone is more likely to be at risk of suicide the next day. 

The study involved 79 adults who reported thoughts of suicide within the previous month, including many who also experienced mental health issues such as depression or post-traumatic stress disorder. The study focused on nighttime smartphone users because the activity can disrupt sleep, which may lead to a higher risk of suicidal thoughts. 

The research team used artificial intelligence to analyze nighttime smartphone behaviors through millions of high-resolution smartphone screenshots. They also collected traditional clinical markers of wellbeing through surveys and check ins. 

They found that at-risk individuals were more likely to have suicidal thoughts the next day if they passively used their phones, such as endlessly scrolling, late at night (from 11 p.m. to 1 a.m.). Yet, the results also showed that if people used their phone in an active way in the middle of the night (from 1 a.m. to 5 a.m.) — such as typing on the keyboard to message someone — they had fewer suicidal thoughts the next day. 

These findings are inline with growing evidence that when people are actively engaged on their smartphones, it may actually help them cope, Jacobucci says.

“We found that late-night messaging was actually beneficial, presumably because it reflects social connection at a vulnerable moment,” he says. 

In contrast, using it passively can have the opposite effect, Ammerman added, because it may heighten suicidal thoughts by fueling unhealthy comparisons, increasing exposure to content that evokes negative emotion, or contributing to sleep disruption.

This small but promising study could make an impact on better understanding the nuances of such a widespread public health crisis as suicide.

“Suicide touches far more people than many realize; about two in five U.S. adults say they know someone who died by suicide,” says Ammerman. "This ripple effect is part of what makes this work feel urgent.”

Harnessing the Power of AI in Research

Jacobucci, Ammerman and their collaborators at University of Notre Dame spent more than a year collecting around 7.5 million smartphone screenshots from individuals in the study to better understand the interplay between  nighttime smartphone behavior and suicidal thoughts and planning. 

“Screenshot capture — every 5 seconds while the phone is in use — lets us understand what’s happening in real time at a finer time scale than self-report alone,” Jacobucci says.

Participants produced an average of 92,000 screenshots, offering a more complete representation of how suicide risk can fluctuate minute to minute.