Non-suicidal self-injury (NSSI) is a highly prevalent behavioral problem among depression adolescent patients that can result in numerous adverse outcomes. This study endeavors to bridge this knowledge gap by creating a comprehensive model that incorporates multiple aspects of NSSI to accurately evaluate its risk in adolescents with depression, thereby enhancing our ability to prevent and address this challenging issue.
Using a cross-sectional design, we recruited 302 adolescents with depressive disorders who visited or were hospitalized at Shandong Mental Health Center from December 2021 to June 2022. The participants completed several self-report questionnaires, including the Chinese version of the Internet Addiction Test, the Pittsburgh Sleep Quality Index questionnaire, the Defeat Scale, the Social Avoidance and Distress Scale and the Children’s Depression Inventory. Logistic regression analysis was performed to identify the diagnostic factors, which were further used to establish clinical risk assessment models. A receiver operating characteristic curve (ROC) to identify the best model. An external validating team was introduced to verify the assessing efficiency.
Based on a logistic regression analysis, three variables have been identified as significant risk factors. Specifically, adolescents with depression who experience low self-esteem, internet use, or suffer from sleep disturbance face an increased risk of NSSI. An integrated risk index for NSSI exhibits excellent accuracy in identifying depressed adolescents at risk of NSSI (area under the curve = 0.86, sensitivity = 0.88, specificity = 0.69). In the validation cohort, the identification performance remains strong (area under the curve = 0.84, sensitivity = 0.72, specificity = 0.81).
This study highlighted the role of self-esteem, internet use and sleep disturbance in the development of NSSI. The risk index diagnosing NSSI onset may help to guide the design and application of novel interventions to minimize this risky behavior in future depressed adolescents.