Teen mental health in the U.S. is reaching a critical point.
According to the CDC, more than 40% of high school students reported feeling persistently sad or hopeless, and nearly 1 in 5 seriously considered suicide.
For a lot of teens, mental health struggles don’t begin with one dramatic moment. They build slowly through pressure and the intense emotional ups and downs of adolescence.
“It’s a biologically more vulnerable period of time,” said Dr. Jonathan Posner, a child psychiatrist at Duke University School of Medicine.
At the same time, experts say many young people struggle for years before they ever see a mental health professional.
Now researchers are using AI to help doctors identify which teens are at risk before a crisis begins.
Doctors say early warning signs can be subtle and there aren’t enough child mental health specialists to screen every child.
“We need to do better at identifying these kids, identifying them early, and getting them in treatment before there’s actually a crisis,” Posner told Ivanhoe.
That’s why researchers at Duke are testing an artificial intelligence model designed to spot early warning signs by analyzing a much broader picture of a teen’s life.
“I think we’re missing the whole picture,” said Dr. Matthew Engelhard, an AI researcher at Duke University.
The model looks at data from more than 11,000 children across the US – including family conflict, social factors and health data. The goal isn’t to replace doctors but to help them see risk sooner.
“With a relatively brief set of questionnaires, the AI tool can actually identify who’s at risk,” Posner explained.
One discovery that surprised even the researchers?
“We expected sleep to emerge as important, but I don’t think we anticipated the degree to which it was important,” Engelhard said.
The system can even estimate whether a teen may develop a mental health condition up to a year before symptoms appear with 75% accuracy –-- which means spotting trouble early could change, or even save, a life.
The Duke research team is now testing the AI tool in clinics to see how well it works outside the lab.
Doctors say primary care physicians often don’t have time to conduct detailed psychiatric assessments, but this tool could automate the process and analyze data in real time, flagging which teens may be at risk during a routine checkup long before a crisis point.