Prediction of suicidal ideation and depression in the general population with subthreshold insomnia using machine learning models

This study leveraged machine learning models to indirectly screen for suicidal ideation (SI) and moderate-to-severe depression in a Slovenian nationwide community sample, specifically focusing on individuals with subthreshold insomnia. The models, developed using socio-demographics, life satisfaction, behavioral changes, and coping strategies, demonstrated robust and consistent predictive performance across both insomnia and non-insomnia groups. The findings highlight the potential for these ML models as feasible, ethical, and efficient tools for early detection in high-risk populations, particularly those presenting with sleep complaints, potentially reducing preventable morbidity and mortality.