Predicting suicidal thoughts and behavior among adolescents using the risk and protective factor framework: A...

Area of Science:

• Adolescent Health
• Machine Learning in Public Health
• Mental Health Research

Background:

• Understanding adolescent suicidal thoughts and behavior (STB) requires identifying interacting risk and protective factors.
• Machine learning offers advanced methods to uncover complex relationships influencing adolescent STB.
• Social determinants of health are crucial in the context of adolescent mental health.

Purpose of the Study:

• To develop a machine learning algorithm for predicting STB in adolescents.
• To identify key risk and protective factors contributing to adolescent STB.
• To utilize a risk and protective factor framework combined with social determinants of health.

Main Methods:

• Utilized a large dataset of over 179,000 high school students' survey responses (2011-2017).
• Integrated demographic data from the American Census Survey.
• Employed interpretable machine learning techniques to identify predictive factors for STB.

Main Results:

• Achieved 91% accuracy in predicting individuals with STB.
• Identified top ten predictive questions across four categories: familial life, drug consumption, demographics, and peer acceptance.
• Highlighted digital harassment, school bullying, and family arguments as leading predictors.

Conclusions:

• Machine learning models significantly improve STB prediction accuracy in adolescents.
• Identified specific risk factors like online harassment and family conflict as critical predictors.
• Findings can inform targeted prevention programs and policy decisions for adolescent mental health.