In Journal of affective disorders ; h5-index 79.0
BACKGROUND : Accurately identifying suicidal high-risk groups and taking intervention measures are important to reduce future suicide risk. In this study, a nomogram technique was used to develop a predictive model of middle school students' suicidality from four aspects: individual characteristics, health risk behaviors, family factors, and school factors.
METHODS : A total of 9338 middle school students were surveyed using stratified cluster sampling method, and subjects were randomly divided into training set (n = 6366) and validation set (n = 2728). In the training set, combined with the results of the lasso regression and random forest, the most important 7 predictors of suicidality were screened and a nomogram was constructed. Discrimination, calibration, clinical applicability, and generalization of the nomograms were assessed by using receiver operating characteristic curves (ROC), calibration curves, decision curve analysis (DCA), and internal validation.
RESULTS : Gender, depression, self-injury, running away from home, parental relationship, relationship with father, and academic stress were found to be significant independent predictors of suicidality. The area under the curve (AUC) of the training set is 0.806, the AUC of the validation data is 0.792, the calibration curve of the nomogram is close to the diagonal, and the clinical decision curve shows that there are more net benefits under the threshold probability of 9-89 %.
CONCLUSION : This study has constructed a good tool for predicting the risk of suicidality among middle school students. It is helpful for school health personnel to assess the risk of suicidality among middle school students, and to identify high-risk groups.
Yan Jie, Liu Yang, Yu Junjie, Liao Lipin, Wang Hong
2023-Feb-17
Machine learning, Middle school students, Nomogram, Suicidality
