Motivational–addictive profiles of nonsuicidal self-injury in Chinese youth: a cluster analysis with validation using external correlates

Background Nonsuicidal self-injury (NSSI) is heterogeneous, yet clinically useful exploratory subgrouping remains limited. We examined whether motivational and addiction-like features could delineate meaningful data-driven NSSI profiles and relate to clinically relevant outcomes. Methods We clustered six motive dimensions of the Ottawa Self-Injury Inventory (OSI; F1–F6) and an addiction-like score in adolescents (N = 311; age range: 12–21 years; mean age = 15.15 years, SD = 2.60). A three-cluster solution was retained based on a multi-criterion evaluation of internal fit indices, cluster-size balance, and clinical interpretability, with the WCSS elbow plot showing a clear inflection at k = 3. Outcomes were tested in age/sex-adjusted models (ordinal frequency and pain via proportional-odds; Patient Health Questionnaire-9[PHQ-9]/Generalized Anxiety Disorder 7[GAD-7] scores were standardized in the full sample [n = 311]). Results Three data-driven profiles emerged: Emotion-regulation (n = 112), Multi-motivated/high-addiction (n = 79), and Lower-severity (n = 120). Versus Lower-severity, both higher-severity clusters showed higher monthly (OR = 3.58 and 2.38) and yearly NSSI frequency (OR = 2.91 and 1.96). Pain perception was lower in the high-addiction cluster (OR = 0.49). PHQ-9 and GAD-7 were higher in the high-addiction cluster, and GAD-7 was also higher in the emotion-regulation cluster. Conclusions Motivational–addictive features identified interpretable, data-driven NSSI profiles associated with frequency, pain, and affective symptoms; these exploratory profiles require replication, longitudinal follow-up, and independent validation before further clinical inferences can be drawn.