AI-ASSISTED SCREENING OF DEPRESSION AND ANXIETY AMONG UNIVERSITY STUDENTS : A CROSS-SECTIONAL STUDY

Abstract

Background: Depression and anxiety represent significant and growing mental health challenges among university students worldwide. Traditional screening methods such as the Patient Health Questionnaire-9 (PHQ-9) and Generalized Anxiety Disorder-7 (GAD-7) remain reliable but are often time-consuming and resource-intensive. Advances in artificial intelligence (AI) offer the potential to augment or automate screening processes, enhancing accessibility and efficiency in academic environments. Objective: This study aimed to evaluate the diagnostic accuracy and agreement of an AI-assisted screening tool against standardized psychometric instruments for detecting depressive and anxiety symptoms among university students. Methods: A total of 370 students were enrolled, with 356 completing the full assessment (response rate: 96.2%). Participants completed the PHQ-9 and GAD-7 scales, and results were compared with AI-generated predictions. Diagnostic performance was assessed using sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and Cohen’s kappa coefficient. Pearson correlation coefficients were used to determine associations between AI and traditional scale scores. Results: The prevalence of clinically significant depression and anxiety (score ≥ 10) was 33.7% and 35.4%, respectively. The AI tool classified 33.1% as probable depression and 36.5% as probable anxiety, demonstrating strong concordance with PHQ-9 and GAD-7 results. Sensitivity and specificity were 89.2% and 91.6% for depression, and 87.6% and 90.1% for anxiety. Substantial agreement was observed (κ = 0.82 and κ = 0.79, p < 0.001). ROC analyses yielded AUCs of 0.94 for depression and 0.92 for anxiety, indicating excellent predictive accuracy. Conclusion: The AI-assisted screening tool demonstrated strong diagnostic validity and agreement with established psychometric scales, supporting its potential as an efficient and reliable adjunct for large-scale mental health screening among university students.

Ähnliche Arbeiten