J Psychiatr Res. 2025 Oct 24;192:142-153. doi: 10.1016/j.jpsychires.2025.10.054. Online ahead of print.
ABSTRACT
BACKGROUND: Non-suicidal self-injury (NSSI) represents a significant public health concern among adolescents diagnosed with mood disorders. Despite its prevalence, predictive models for NSSI based on social ecological factors remain underdeveloped. Therefore, this study aims to develop and validate a risk prediction model that facilitates early screening and intervention for adolescents at high risk of engaging in NSSI.
METHODS: A cohort of 467 adolescents with mood disorders was divided into a training set and an internal validation set. An external dataset of 755 general adolescents was employed for external validation. To identify optimal predictors of NSSI, random forest and LASSO regression models were utilized. The selected variables were subsequently used to construct a comprehensive prediction model. The model’s discrimination, calibration, and clinical utility were validated using areas under the curve (AUC), calibration plots, and decision curve analysis (DCA).
RESULTS: Seven key predictors for NSSI were identified: sex, sleep problems, psychological suffering, problematic mobile phone use, self-evaluation, perceived discrimination at school, and learning burnout. The predictive model demonstrated strong discrimination capabilities in both internal and external validation phases. In both datasets, the calibration curves exhibited bias correction and a clear resemblance to the ideal line. The DCA curve indicated that the model provides a net benefit across a broad range in both internal and external validation.
CONCLUSIONS: This study reports the successful development and external validation of a comprehensive risk prediction model for NSSI among adolescents with mood disorders. The model effectively integrates seven critical predictors spanning individual, family, and school domains. Presented as a nomogram, this predictive instrument serves as a practical and effective tool for clinicians and practitioners to identify high-risk individuals, facilitating timely interventions and significantly contributing to NSSI prevention efforts.
PMID:41161223 | DOI:10.1016/j.jpsychires.2025.10.054
