Science & Technology, Australia (Commonwealth Union) – The use artificial intelligence (AI) continues to expand into variety of areas that include health and science. The benefit of AI in processing vast amounts of data continues to add new benefits.
New research conducted by The University of New South Wales (UNSW) reveals that AI can play a crucial role in pinpointing risk factors associated with suicide and self-harm.
In Australia, suicide stands as the leading cause of death among adolescents, with self-harm affecting approximately 18% of individuals aged 14 to 17. Regrettably, these issues have become more prevalent in this age group over the past decade.
In healthcare settings, like hospitals, clinicians typically assess the risk of suicide and self-harm when young individuals exhibit potential self-destructive behavior. However, current risk assessment methods, such as examining past suicide attempts, often prove to be unreliable, as they overlook numerous other potential risk factors. Additionally, adolescents who are not within these healthcare settings may go unnoticed.
The field of mental health is increasingly turning to artificial intelligence (AI) to identify individuals at risk. Machine learning (ML) models have the capacity to analyze extensive patient data, recognizing potential risk factors and gauging their predictive capabilities concerning mental health issues, including suicide and self-harm attempts.
Researchers hailing from UNSW, the Ingham Institute for Applied Medical Research, and South Western Sydney Local Health District (SWSLHD) have developed ML models designed to forecast the likelihood of suicide and self-harm attempts in adolescents. These models outperform conventional approaches, as they rely on more than just a history of suicide or self-harm attempts as risk factors.
The latest research findings have been unveiled recently in the Psychiatry Research journal.
According to Dr. Daniel Lin, a senior author and a psychiatrist and mental health researcher associated with UNSW, the Ingham Institute, and SWSLHD, clinicians often encounter the challenge of processing vast amounts of information that can exceed their capacity. To address this, they have turned to harnessing the power of machine learning algorithms.
To conduct their study, the team of researchers utilized data from the Longitudinal Study of Australian Children, an ongoing data collection effort that has been tracking a diverse range of information about children throughout Australia since 2004. Their analysis focused on 2,809 participants, who were divided into two age groups: 14-15 years and 16-17 years. The data used in their study was sourced from questionnaires completed by the children themselves, their caregivers, and their school teachers.
Among the 2,809 participants, it was revealed that 10.5 percent had reported engaging in self-harming behaviors, while 5.2 percent had reported attempting suicide at least once within the past 12 months. Dr. Lin emphasizes that these figures likely underestimate the true prevalence of such behaviors, suggesting that the actual proportions are likely higher.
From the extensive dataset, the researchers pinpointed over 4,000 potential risk factors encompassing various domains such as mental health, physical well-being, interpersonal relationships, and the school and home environment. Employing an advanced machine learning technique known as the random forest classification algorithm, they sought to determine which risk factors observed at the age of 14-15 exhibited the highest predictive power for instances of suicide and self-harm attempts at the age of 16-17.
In the context of suicide and self-harm, the most critical risk factors included feelings of depression, emotional and behavioral challenges, self-perceptions, as well as dynamics within the school and family environments. Additionally, there were distinct factors specific to either suicidal tendencies or self-harming behaviors.
“A unique predictor of suicide was lack of self-efficacy, when someone feels a lack of control over their environment and their future. And a unique predictor of self-harm was lack of emotional regulation,” explained Dr Lin.
“It was surprising to us to see that previous attempts were not among the top risk factors.”
