BYU researchers create an algorithm that can predict teen suicidal behavior

*CONTENT WARNING: This article contains mentions of teenage suicidality.*

PROVO (ABC4) – Algorithm can be a scary word, admits Quinn Snell, professor of computer science at Brigham Young University.

The term, which has since achieved commonplace status in the modern lexicon, may conjure up imaginings of intrusive data analysis by artificial intelligence-based supercomputers that can understand human nature better than we as a species understand. ourselves.

Snell acknowledges that the phrase can be frightening to many, but when used correctly, algorithms can be used to create impactful and positive change in human society.

“When we talk about algorithm, it can turn people off and make them nervous with artificial intelligence and all the hype around it. But what we’re really talking about is that the data tells us a story” “And we’re just trying to figure out what this story is, and all the program does is try to figure out what this story is,” Quinn tells

The story that Snell and his academic cohorts, Associate Dean of the College of Life Sciences Michael Barnes and Professor of Public Health Carl Hanson, are uncovering from data from their latest research involving the use of computer learning ; what are the main predictors of suicidal thoughts and behaviors in adolescents.

Using a massive amount of data, including surveys of nearly 180,000 Utah middle and high school students that consisted of over 300 questions and included over 8,000 items of demographic detail, the trio, with the help of researchers from Johns Hopkins and Harvard, found an algorithm that could predict harmful thoughts and behaviors with an astonishing 91% accuracy rate.

Hanson believes the results could and should make a major difference in potential actions and public health changes that could ultimately save lives.

“We believe that certain things in our particular analysis have been shown to be major risk factors for suicidal thoughts and behaviors,” Hanson said, noting that suicide is the second leading cause of death among American adolescents. “And those are the important things that we think could begin to address the ultimate problem of suicide in our communities with teenagers.”

Using the data, the team was able to identify the most influential and predictable risk factors for suicidal thoughts and behaviors. The factors could be summarized in two large “stacks” of interactions for young people; peer connections and home connections.

While two of the main risk factors that predicted harmful behavior with the greatest accuracy were arguments at home and bullying at school, the most predictable factor for self-harming thoughts or behaviors was bullying or bullying. online bullying.

Although the researchers were already aware of the dangerousness of these three environments, they were surprised at the severity and predictability of online bullying, in-person bullying, and argued family life to bring a teenager to consider or act out desperate measures.

“I think we’ve known for a long time that there are negatives in the online world,” Barnes says. “And obviously there are positives and there are good things that go into that, but the ugly side comes out in times like this. Our results clearly show a sort of glass-half-empty approach to the online world. »

This shows that relationship health, whether online, as many young people experience through social media and other platforms, may be as important to a person’s overall well-being as any other aspect of health. Barnes imagines a future where data that illustrates this importance leads to an evolution of what healthcare might look like in the future.

“What would a health clinic look like if we focused on relationships, not just physiology? It’s as important as not only the food we eat, but also the people we are lucky enough to be able to enjoy the food with,” he projects.

And while technology may have opened the door to a major problem in the form of online bullying, innovations like the computer science that found the algorithm may also open the door to the solution. The convergence of public health, life sciences and computer learning technologies such as BYU’s published research on teen suicide prevention illustrates how different disciplines can work in harmony to create real-world solutions.

Computing can have a big impact on a better understanding of how to heal and treat humans living behind screens and keyboards.

“We would like to continue working with this data, not just on this, but also looking at other things with this data,” Snell says. “Computational health sciences are our focus, those that examine the ways in which computing and data analytics can affect our public health.”

If you are considering suicide, help is available. The National Suicide Prevention Lifeline is available 24/7 at 1-800-273-8255 or online at