Associations Between Psychological Factors, Routine Hematological and Biochemical Indicators, and Non-Suicidal Self-Injury Among University Students: An Exploratory Cross-Sectional Study

Background: Non-suicidal self-injury (NSSI) is a major mental health concern among university students. Although psychological risk factors for NSSI have been extensively studied, the potential contribution of routinely available hematological and biochemical indicators remains unclear. This study examined their associations with NSSI and their incremental value beyond psychological factors. Methods: This cross-sectional study included 1,661 university students who completed campus psychological screening and routine health examinations. NSSI status was defined as an NSSI total score greater than zero. Psychological variables included depressive symptoms, anxiety symptoms, sleep problems, impulsivity, attentional control, cognitive emotion regulation strategies, and problematic internet use. Associations between routine indicators and NSSI severity were examined using Spearman correlations, while logistic regression models evaluated associations with NSSI status after adjustment for demographic and psychological factors. Five-fold cross-validation was used to compare predictive performance across demographic, routine indicator, psychological, and combined models. Results: Among the 1,661 participants, 244 (14.69%) were classified as having NSSI. Several routine indicators, including AST, ALT, red blood cell count, uric acid, hemoglobin, and hematocrit, showed weak correlations with NSSI total scores after FDR correction. However, no routine hematological or biochemical indicator remained significant after false discovery rate (FDR) correction in either adjusted logistic regression model. The psychological model showed substantially better predictive performance than demographic or routine indicator models, and adding routine indicators did not improve prediction (AUC: 0.796 vs. 0.787). Conclusion: Routine hematological and biochemical indicators showed weak associations with NSSI severity but provided limited incremental value beyond psychological factors for identifying NSSI among university students. Psychological assessment remains an important component of NSSI screening, whereas routine laboratory data did not show additional incremental value in this sample. Keywords: non-suicidal self-injury, university students, hematological indicators, biochemical indicators, psychological screening, predictive modeling