ABSTRACT
Introduction
Non-suicidal self-injury (NSSI) and suicidal ideation (SI) are prevalent and co-occurring among adolescents, serving as critical predictors of suicide. This study aimed to develop a predictive model and nomogram for suicide attempts (SA) in Chinese adolescents with mood disorders exhibiting NSSI and SI.
Methods
Data were collected from 134 participants. Predictors were selected via LASSO regression from data collected using the Self-Injurious Thoughts and Behaviors Interview-Revised and self-report scales, followed by multivariate logistic regression to build the nomogram. Model performance was assessed through discriminatory ability, calibration curves, and clinical decision analysis.
Results
Adolescents with SA history had fewer education years, higher prevalence and future likelihood of self-injurious thoughts and behaviors, earlier NSSI onset, more frequent and severe NSSI, and more intense and persistent SI compared to those without SA. Three key predictors for SA were identified: NSSI emotion regulation scores, average SI persistence duration, and history of interrupted attempts. The developed nomogram exhibited robust predictive accuracy with an AUC of 0.756.
Discussion
This study presents a predictive model for suicide risk in adolescents with mood disorders exhibiting NSSI and SI. The model demonstrates high predictive accuracy and clinical applicability, offering a practical tool for clinicians to prioritize high-risk cases and guide personalized interventions.