A Machine Learning-Based Method for Developing the Chinese Symptom Checklist-11 (CSCL-11).

Area of Science:

• Psychiatry and Psychology
• Machine Learning in Healthcare
• Psychometric Instrument Development

Background:

• The Chinese Symptom Checklist-90 (SCL-90) is lengthy, leading to low compliance.
• A shorter, efficient screening tool is needed for the Chinese general population.

Purpose of the Study:

• To develop a concise psychopathological screening instrument using a subset of the Chinese SCL-90 items.
• To identify individuals at high risk for psychological disorders using machine learning.

Main Methods:

• Variable clustering was applied to select 11 representative items from the SCL-90.
• Factor analysis was used to validate the psychometric properties of the new scale.
• Machine learning models were employed to assess predictive performance and identify high-risk individuals.

Main Results:

• The resulting 11-item scale, Chinese Symptom Checklist-11 (CSCL-11), showed high internal consistency (Cronbach's α = 0.84).
• The CSCL-11 demonstrated strong predictive accuracy for Global Severity Index (GSI) scores (R² = 0.92) and identified high-risk individuals with 96% accuracy.
• The CSCL-11 outperformed other shorter scales (SCL-14, SCL-K11, SCL-K-9) in predicting GSI scores and identifying high-risk groups.

Conclusions:

• The CSCL-11 is a concise and effective preliminary screening tool for psychopathology in the Chinese general population.
• It retains essential information from the SCL-90 while improving completion rates and efficiency.
• The CSCL-11 offers a valuable alternative for large-scale psychological screening.