Related Experiment Video
Updated: Nov 10, 2025
10:39
Qualitative and Quantitative Validation of Tools with Rating Scales Aimed at Assessing the Quality of University Service-Learning
Published on: August 29, 2025
385
Which PHQ-9 Items Can Effectively Screen for Suicide? Machine Learning Approaches.
Sunhae Kim1, Hye-Kyung Lee2, Kounseok Lee1
1Department of Psychiatry, Hanyang University Medical Center, Seoul 04763, Korea.
Summary
Machine learning algorithms effectively use the Patient Health Questionnaire-9 (PHQ-9) to accurately screen for suicidal ideation in college students. This tool shows high accuracy, aiding early detection in primary care settings.
Area of Science:
- Psychiatry
- Data Science
- Public Health
Background:
- The Patient Health Questionnaire-9 (PHQ-9) is a widely used depression screening tool in primary care.
- Its efficacy in evaluating suicidal ideation requires further investigation.
Purpose of the Study:
- To assess the accuracy of the PHQ-9 in identifying suicidal ideation among college students.
- To compare the performance of machine learning algorithms using PHQ-9 data for suicide risk assessment.
Main Methods:
- Analysis of 8760 completed PHQ-9 questionnaires from college students.
- Evaluation of PHQ-9 against PHQ-2, PHQ-8, and PHQ-10 scoring variations.
- Utilized machine learning algorithms (k-nearest neighbors, LDA, random forest) with suicidal ideation as the dependent variable, validated by the Mini-International Neuropsychiatric Interview suicidality module.
Main Results:
- Random forest model using PHQ-9 demonstrated excellent performance with an AUC of 0.841 and 94.3% accuracy.
- Achieved high positive predictive value (84.95%) and negative predictive value (95.54%) for detecting suicidal ideation.
Conclusions:
- Machine learning algorithms combined with the PHQ-9 reliably and accurately screen for suicidal ideation.
- The findings support the integration of PHQ-9 and ML for enhanced suicide risk detection in primary care.
