Absolute Risk Prediction for Cannabis Use Disorder in Adolescence and Early Adulthood Using Bayesian Machine Learning

Area of Science:

• Addiction research
• Machine learning in public health
• Adolescent psychology

Background:

• Substance use disorders (SUD) are a significant public health issue in the US.
• Adolescent substance use can lead to adult SUDs, necessitating early intervention.
• Cannabis use disorder (CUD) is a growing concern among adolescents and young adults.

Purpose of the Study:

• To develop and validate an absolute risk prediction model for CUD in adolescents and young adults.
• To identify key risk factors associated with the development of CUD.
• To provide a tool for clinicians to assess individual CUD risk.

Main Methods:

• A Bayesian machine learning model was trained using data from the National Longitudinal Study of Adolescent to Adult Health.
• The model predicts personalized CUD absolute risk for cannabis-using adolescents and young adults.
• Performance was evaluated using five-fold cross-validation (AUC, E/O) and independent validation datasets.

Main Results:

• The model identified five key risk factors: biological sex, delinquency, conscientiousness, neuroticism, and openness.
• The model demonstrated good discrimination and calibration, with AUC values ranging from 0.64 to 0.75 and E/O values close to 1 across datasets.
• The model accurately predicts CUD risk within 5 years of first cannabis use.

Conclusions:

• The developed model can assist clinicians in identifying adolescents and young adults at high risk for CUD.
• Early risk assessment enables timely and targeted clinical interventions.
• This predictive tool supports public health efforts to mitigate the progression of cannabis use disorder.