Building and evaluating suicide attempt prediction models using risk factors.

Area of Science:

• Psychiatry
• Public Health
• Data Science

Background:

• Suicide attempts pose a significant public health challenge.
• Accurate prediction models are crucial for early detection and intervention.

Purpose of the Study:

• To identify risk factors for suicide attempts.
• To build and evaluate the performance of suicide attempt prediction models.

Main Methods:

• Secondary data analysis of 11,671 adults aged 19 years and older.
• Multiple logistic regression to identify risk factors.
• Performance analysis included calibration, discrimination, and clinical usefulness.

Main Results:

• Identified risk factors: suicide plan and suicidal ideation (males); suicide plan and depression diagnosis (females).
• Prediction models demonstrated good calibration and discrimination (AUC > 0.90).
• High sensitivity achieved: 90.9% (males) and 82.4% (females) at 0.5% cutoff.

Conclusions:

• Developed suicide attempt prediction models show acceptable performance.
• Models can effectively assess risk and identify high-risk populations early.
• Facilitates targeted community interventions with limited resources.