Concise multi-class anxiety disorder risk assessment: A novel advanced machine learning approach

Area of Science:

• Psychiatry and Mental Health
• Computational Neuroscience
• Health Informatics

Background:

• Traditional anxiety disorder assessments like the Depression Anxiety Stress Scales-42 (DASS-42) are time-consuming for patients and clinicians.
• Efficient and accurate mental health screening is vital for timely intervention and improved patient outcomes.
• The need for rapid, scalable anxiety disorder risk assessment tools is significant, particularly for underserved populations.

Purpose of the Study:

• To develop a concise anxiety disorder scale using advanced machine learning (ML) techniques.
• To maintain high validity while significantly reducing the number of questions required for risk assessment.
• To create an accessible online tool for rapid anxiety self-screening and clinical use.

Main Methods:

• Applied advanced machine learning algorithms and feature selection to the DASS-42 questionnaire.
• Developed and validated a novel, condensed anxiety risk assessment model.
• Integrated the optimized scale into a user-friendly online screening tool.

Main Results:

• A concise anxiety disorder scale was successfully created, maintaining high predictive validity.
• The optimized model requires fewer questions than traditional surveys for effective anxiety risk prediction.
• The online tool provides efficient and accurate anxiety risk assessment for self-screening and clinical settings.

Conclusions:

• Machine learning can effectively create practical and accessible mental health assessment tools.
• This innovation facilitates faster and earlier anxiety disorder detection and intervention.
• The study demonstrates significant societal implications for improving mental well-being through accessible technology.