17.4 percent (or 1 in 6) of US children aged two to eight have a diagnosed mental, behavioral, or developmental disorder, according to the CDC. Additionally, suicide attempts and deaths among children have increased in the US over the past decade, and suicide is now the eighth leading cause of death in children at the tender ages of 5–11. And the situation has only worsened since the pandemic. According to an article in USA Today, Dr. Heather Huszti, chief psychologist at Children’s Health of Orange County, states, “The number of mental health emergencies involving children was a fire before, and the pandemic was a can of gas we poured on that fire. I’ve practiced over 30 years (and) I’ve never seen anything like this.”
So how can doctors and mental health professionals combat these horrific statistics and suicide rates? According to Mental Health America, early detection is known to help alleviate mental health disorders in early childhood. And different forms of intervention can help lessen or prevent mental illness from spilling over into adulthood. With this in mind, doctors from Cincinnati Children’s Hospital Medical Center are using the world’s second-most powerful supercomputer and artificial intelligence (AI) to assess mental health risk factors in children. The team hopes that by collecting and computing mental health trajectories, more advanced diagnosis and treatment will follow. Below, we take an in-depth look at how this ambitious project may save lives and revolutionize pediatric mental health.
A Massive Undertaking
The Cincinnati Children’s Hospital mental health project is tremendous in both size and scope. It will involve more than 25 leading scientists at nine different research divisions within Cincinnati Children’s Hospital and collaborators at the University of Cincinnati, the University of Colorado, and Oak Ridge National Laboratory, the owner and operator of the Summit supercomputer. According to Forbes, “Built by IBM, Summit was the first supercomputer to hit exaflop speed: a quintillion operations per second.” So what does this mean in layman’s terms? The supercomputer trains the AI models used to assess mental health risk factors more than 17,000 times faster than an average laptop.
Why the advanced technology? According to the Cincinnati Children’s Hospital science blog, “mental illness is extraordinarily complex.” With this in mind, the project’s team will gather large amounts of data from several sources that researchers already know to play a factor in influencing a growing child’s mental health. Such information includes personal medical records, economic disruption, family issues, substance abuse, bullying, systemic racism, and other factors. With such an intense amount of data – calculating them all is challenging to say the least. Enter the Summit computer. Summit’s mission was designed to provide scientists with “computing power to solve challenges in energy, AI, human health, and other research areas that were simply out of reach until now,” according to the Oak Ridge National Laboratory’s website. “We are excited to leverage Oak Ridge National Laboratory’s (ORNL) world-class leadership computing capabilities to impact the future of pediatric mental health,” ORNL principal investigator Greeshma Agasthya tells the Cincinnati Children’s Hospital science blog. “Improving pediatric mental health outcomes is a major challenge. We are dedicated to using our expertise to capitalize on the wealth of information available in health records, medical notes, and medical images while ensuring the privacy of the data.”
The Project’s Challenges & Its “Heartbreaking” Modeling Data
According to Forbes, one of the significant challenges in the project is identifying natural language. More specifically, the processing to accept what children are saying, understanding it, and using that as part of the input for the risk assessment models. Another issue is staying up-to-date; the project software continually updates itself by reading PubMed and Medline articles to pinpoint research findings that might be important to update models and highlight them to a human. The project aims to create near real-time modeling for the early identification of at-risk children, particularly for depression and suicidal ideation. First, however, medical and mental health professionals need to know what they are looking for to accomplish this task – or the signs of suicidal thinking. However, the key resources used in this type of modeling are heartbreaking and challenging to examine – but critical.
According to John Pestian, Ph.D., MBA, an expert in building neuropsychiatric artificial intelligence algorithms and co-principal investigator for the project, “They’re all built off this massive collection of suicide notes that I collected and then built natural language models off of those,” he tells Forbes. “And these are notes that people wrote just before they died by suicide.” Pestian and colleagues took those notes and designed questions to ask kids, like “do you have secrets” and “are you angry.” The answers are part of what the AI model will use to determine if kids are at high, medium, or low risk of depression or other mental health challenges. However, the answers are not the only critical data used. The computer also relies on how individuals talk, how many pauses they take between sentences and the facial expressions they make as they talk.
Need for a Human Touch
The AI in this enormous project is vital, but it is not a decision tool. This is where human involvement is essential. Pestian tells Endominance, “When it comes to caring for mental illness, it’s unlikely that machines will ever totally replace humans. It is just too complex because it combines both biological data and thought data. The role of the machine is to identify all the important complex data and present it in a way that the caregivers can understand. For the near term, artificial intelligence methods are not sophisticated enough to fully understand all the facets of mental illness. So they act as decision support and not decision making.” Pestian also tells Forbes, “Wherever decisions for at-risk assessments lie, the eventual intervention will also have a human touch. So physicians, mental health professionals, and school counselors are critical.”
The Economic Impact & a Call to Arms
The emotional and psychological toll of mental health issues in children is evident. But what about the economic impacts? Pestian tells the Cincinnati Children’s Hospital science blog, “We have found studies from other researchers that estimate the economic impact of anxiety at $42 billion a year, depression at $71 billion a year, and suicidal ideation at $93.5 billion a year.” In truth, all of these horrific statistics on children’s mental health and suicide should be a call to arms for every pediatric medical and mental health professional. The need for better diagnosis, intervention and treatment is paramount. Full stop.
“If we can identify this early, we can treat for and alleviate almost 50 percent of the mental illness that goes into adulthood, so catching it young, catching it early, and giving care is a very important part,” Pestian tells the podcast TechFirst. Pestian foresees having a trained model within a year but also thinks the foundation will need to invest another few years in working out any issues before being available at scale.
