Predicting adolescent problem use of marijuana: development and testing of a Bayesian model

Area of Science:

• Psychology
• Public Health
• Computer Science

Background:

• Problem marijuana use among teenagers is a significant public health concern.
• Existing risk assessment tools may lack comprehensive predictive power or require extensive data.
• Computer-based support systems offer a scalable platform for adolescent interventions.

Purpose of the Study:

• To develop and validate a novel risk assessment index for problem marijuana use in adolescents.
• To integrate this index into a computer-based support system for guiding teenagers.
• To enhance marijuana-related behavioral interventions through improved risk prediction.

Main Methods:

• Utilized Bayesian decision theory for index development, leveraging expert judgments.
• Developed a computer-based support system incorporating the risk assessment index.
• Validated the index against an independent panel's risk ratings and longitudinal self-reported marijuana use data.

Main Results:

• The developed risk assessment index accurately predicted an independent panel's assessment of teenager risk.
• The index demonstrated predictive validity for 10th-grade marijuana use based on 7th-grade data.
• Bayesian decision theory enabled index creation without a large empirical database.

Conclusions:

• The risk assessment index is a viable tool for identifying teenagers at risk for problem marijuana use.
• The index can be effectively integrated into computer-based systems to support behavioral change.
• This approach offers a flexible and data-efficient method for future risk assessment tool development.