Related Experiment Video
Updated: Nov 29, 2025
07:30
Brain Morphology of Cannabis Users With or Without Psychosis: A Pilot MRI Study
Published on: August 18, 2020
7.2K
A preliminary risk prediction model for cannabis use disorder
Rajapaksha Mudalige Dhanushka S Rajapaksha1, Ryan Hammonds2, Francesca Filbey2
1Department of Mathematical Sciences, University of Texas at Dallas, Richardson, TX, USA.
Preventive Medicine Reports
|November 18, 2020
Summary
A new model can predict the risk of developing cannabis use disorder (CUD) by analyzing personal risk factors. This tool helps identify individuals needing early intervention for CUD.
Area of Science:
- Psychiatry and Behavioral Science
- Data Science and Machine Learning
Background:
- Cannabis legalization is increasing cannabis use disorder (CUD) prevalence.
- Predicting individual CUD risk is crucial for early intervention.
- No existing models quantify CUD risk based on personal factors.
Purpose of the Study:
- Develop a preliminary model to predict quantitative risk of CUD.
- Identify stable personal risk factors for CUD development.
Main Methods:
- Utilized statistical and machine learning classification techniques.
- Applied leave-one-out cross-validation for model performance evaluation.
- Developed a LASSO logistic regression model using seven key risk factors.
Main Results:
- The final model includes age, initial smoking enjoyment, impulsivity, and personality traits (neuroticism, openness, conscientiousness).
- Achieved an overall accuracy of 0.66 and an AUC of 0.65.
- Identified key predictors for quantitative CUD risk assessment.
Conclusions:
- A preliminary relative risk model for predicting CUD has been developed.
- This model can identify high-risk cannabis users for early intervention.
- Further research can refine CUD risk prediction models.
