中学生抑郁症状预测模型的构建和验证

Establishment and validation of depressive symptom predictive model in middle school students

目的调查中学生抑郁症状的影响因素,并构建和验证预测模型,为提供风险评估工具,有效筛查中学生抑郁症状提供参考。方法分别于2021年9—10月和2022年9—10月,采用多阶段整群抽样方法从广东省某市抽取2376和4344名中学生进行体格检查和问卷调查,并构建中学生抑郁症状的列线图预测模型。问卷调查主要采用学生健康状况及影响因素调查表(中学版)、流调中心抑郁量表测定中学生的生活行为方式和抑郁症状。结果2021年中学生的抑郁症状检出率为23.3%。多因素Logistic回归分析结果显示,早餐不规律(OR=2.64),遭受校园欺凌(OR=4.28),被家长打骂(OR=2.86),移动电子设备使用时间长(OR=1.08)和静坐时间长(OR=1.05)与中学生抑郁症状的发生均呈正相关;睡眠时间长(OR=0.78)和户外活动时间1~<2,2~<3和≥3 h/d(与<1 h/d相比,OR值分别为0.63,0.61,0.49)与中学生抑郁症状的发生均呈负相关(P值均<0.05)。多因素Logistic回归分析有统计学意义的7个预测因子构建列线图,列线图AUC为0.77,经内外部验证,区分度和可靠性良好。结论列线图预测模型可以作为有效评估中学生抑郁症状风险的工具。中学生的生活行为、饮食行为和伤害行为在抑郁症状形成中具有重要作用,应当重视这些行为因素对中学生心理健康的影响。

Objective To investigate the influencing factors of depressive symptoms and to construct and verify the prediction model of depressive symptoms in middle school students,so as to provide risk assessment tools for effectively screening depressive symptom.Methods Physical examination and questionnaire survey were conducted among middle school students in one city in Guangdong Province from September to October in 2021(n=2376)and from September to October in 2022(n=4344)by a multistage cluster sampling method,and a nomographic prediction model of depressive symptoms in middle school student was constructed.The questionnaire survey was conducted using the student health status and influencing factors questionnaire(secondary school version)and the Center for Epidemiological Studies Depression Scale(CES-D)to measure the lifestyle and depressive symptom of middle school students.Results The detection rate of depressive symptoms in 2021 was 23.3%.Multivariate Logistic regression analysis showed that irregular breakfast(OR=2.64),school bullying(OR=4.28),being beaten by parents(OR=2.86),using mobile devices for a long time(OR=1.08)and sitting for a long time(OR=1.05)were positively related to depressive symptoms in middle school students(P<0.05).Long sleep duration(OR=0.78)and outdoor activity durations of 1-<2,2-<3 and≥3 h/d(compared with<1 h/d,OR=0.63,0.61,0.49)were negatively related to depressive symptoms in middle school students(P<0.05).Multivariate Logistic regression analysis showed that 7 statistically signifucant predictive factors constructed a nomogram,and the AUC of the nomogram was 0.77,which had been verified internally and externally with good differentiation and reliability.Conclusions The nomogram prediction model of depressive symptoms provides a convenient and effective risk assessment tool for depressive symptoms among middle school students.The life behavior,diet behavior and injury behavior of middle school students play an important role in the formation of depressive symptoms.It should pay attention to the impact of the behavioral factors on the mental health of middle school students.