Study on the prediction model of non-suicidal self-injury behavior risk during hospitalization for adolescent inpatients with depression based on medical data.

Abstract

Purpose

To develop a predictive model for identifying risk factors of non-suicidal self-injury (NSSI) during hospitalization in adolescents. By analyzing 1242 inpatient records, we explored NSSI risk factors in depressed adolescents and established a clinical predictive nomogram.

Methods

We collected electronic medical records from the First Affiliated Hospital of Zhengzhou University from January 2021 to May 2023. The least absolute shrinkage and selection operator (LASSO) regression with tenfold cross-validation was used for variable selection. Multivariable logistic regression was then applied to build the predictive model. A nomogram was developed based on the selected variables and validated using a calibration plot, receiver operating characteristic curve (ROC), and decision curve analysis (DCA). External validation was also performed.

Results

Six predictors were identified: sex, self-injury within 1 month before hospitalization, current course, history of attempted suicide, with suicide idea, and history of self-injury. The nomogram showed satisfactory discrimination in both the training (AUC 0.927; 95% CI: 0.844-0.905) and validation (AUC 0.907; 95% CI: 0.879-0.902) sets. Decision curve analysis(DCA) indicated clinical utility when the risk threshold was between 15% and 83%, with external validation confirming this range as 17% to 80%.

Conclusion

We developed a nomogram to predict NSSI risk in hospitalized adolescent inpatients with depression. The nomogram demonstrated favorable calibration and discrimination, aiding clinicians in identifying at-risk inpatients and facilitating timely interventions, providing a reference for future prevention.