Construction of a prediction model for comorbidity of depressive and anxiety symptoms among middle school students
10.19485/j.cnki.issn2096-5087.2026.04.010
- Author:
LI Huiyu
;
YANG Yunjuan
- Publication Type:Journal Article
- Keywords:
depressive symptoms;
anxiety symptoms;
comorbidity;
middle school students;
random forest model;
prediction
- From:
Journal of Preventive Medicine
2026;38(4):367-371
- CountryChina
- Language:Chinese
-
Abstract:
Objective :To construct a prediction model for the comorbidity of depressive and anxiety symptoms among middle school students, so as to enable early identification of these symptoms in this population.
Methods:From June to August 2025, a multi-stage stratified random cluster sampling method was used to select students from 8 junior high schools and 8 senior high schools across 4 cities in Yunnan Province. Self-administered questionnaires were used to collect demographic information, including age and gender. The Center for Epidemiologic Studies Depression Scale, the Generalized Anxiety Disorder scale, and the Adolescent Self-Rating Life Events Scale were used to assess depressive symptoms, anxiety symptoms, and life stress caused by negative life events in the past year, respectively. A multinomial logistic regression model was employed to screen predictive factors, followed by the development of a random forest model to predict comorbidity of depressive and anxiety symptoms among middle school students. Model performance was evaluated using receiver operating characteristic (ROC) curves and calibration plots.
Results:A total of 1 173 middle school students were surveyed, including 569 males (48.51%) and 604 females (51.49%). There were 616 junior high school students (52.51%) and 557 senior high school students (47.49%). The comorbidity of depressive and anxiety symptoms was detected in 238 students, with a detection rate of 20.29%. Multivariable logistic regression analysis revealed that gender (female, OR=1.625, 95%CI: 1.118-2.363), residence (rural area, OR=1.845, 95%CI: 1.202-2.832), left-behind experience (OR=1.665, 95%CI: 1.069-2.595), parenting style (authoritative type, OR=0.301, 95%CI: 0.164-0.554), electronic device usage duration (3-<5 h/d, OR=2.442, 95%CI: 1.242-4.803), sleep deprivation (OR=2.316, 95%CI: 1.574-3.407), and negative life stressors (OR=1.110, 95%CI: 1.093-1.128) were significant risk predictors for comorbidity of depressive and anxiety symptoms among middle school students. The random forest model analysis indicated that the top five variables in terms of importance ranking were negative life stress, residence, left-behind experience, sleep deprivation, and electronic device usage duration. The area under the ROC curve, sensitivity, and specificity were 0.834, 0.731, and 0.792 for the training set, and 0.795, 0.732, and 0.771 for the validation set, respectively. The calibration curves showed good model fit for both the training and validation sets (both P>0.05).
Conclusions:The comorbidity of depressive and anxiety symptoms among middle school students in this survey is primarily associated with negative life stress, residence, left-behind experience, sleep, and electronic device usage duration. The constructed random forest model demonstrates certain predictive value.
- Full text:2026110916333652929中学生抑郁焦虑症状共患的预测模型构建.pdf
