基于决策树算法构建中青年重度抑郁症患者发生非自杀性自伤行为风险的预测模型  

Construction of a decision tree algorithm to predict the risk of non-suicidal self-injurious behavior in young and middle-aged patients with major depressive disorder

作　　者：刘浩 周小艳[1] 向桢玉 薛毅 吴清培[1] 黄雪萍[1] LIU Hao;ZHOU Xiaoyan;XIANG Zhenyu;XUE Yi;WU Qingpei;HUANG Xueping(Department of Psychology,Chongqing Mental Health Center,Chongqing 401147,China)

机构地区：[1]重庆市精神卫生中心心理一科,重庆401147

出　　处：《重庆医学》2024年第10期1484-1490,共7页Chongqing medicine

基　　金：重庆市科卫联合医学科研项目(2019MSXM014)。

摘　　要：目的构建基于决策树算法的中青年重度抑郁症患者发生非自杀性自伤(NSSI)行为风险的预测模型。方法选取2021年1月至2023年6月重庆市精神卫生中心收治的中青年重度抑郁症患者作为研究对象,收集患者的临床资料,根据是否发生NSSI将患者分为NSSI组(n=72)和非NSSI组(n=82)。采用单因素和多因素logistic回归分析中青年重度抑郁症患者发生NSSI的危险因素,基于logistic回归分析结果和卡方自动交互检测法(CHAID)建立相关决策树预测模型,受试者工作特征(ROC)曲线评估模型的预测价值。结果154例中青年重度抑郁症患者NSSI的发生率为46.8%。单因素分析结果显示,两组性别、年龄、文化程度、职业、家庭组合方式、是否为独生子、恋爱状况、生育状况、主要居住地、家庭经济来源、住房条件比较差异无统计学意义(P>0.05);居住情况、家庭经济情况、抑郁病程、社会支持评定量表(SSRS)评分、家庭支持自评量表(PSS-Fa)评分和挫败量表(DS)评分比较差异有统计学意义(P<0.05)。多因素logistic回归分析显示,家庭经济水平较低、抑郁病程较长、SSRS评分<20分、PSS-Fa评分<6分和DS评分≥47.2分是中青年重度抑郁症患者发生NSSI的独立危险因素(P<0.05)。构建的决策树模型共5层,10个节点,模型选择了家庭经济情况、抑郁病程、SSRS评分和PSS-Fa评分4个指标作为模型的节点,其中PSS-Fa评分是最重要的预测因子。ROC曲线分析显示,该模型的曲线下面积(AUC)为0.881(95%CI:0.844~0.918)。结论中青年重度抑郁症患者发生NSSI的影响因素较多,基于影响因素构建的决策树模型对中青年重度抑郁症患者发生NSSI行为风险具有较高的预测价值。Objective To construct a predictive model for the risk of non-suicidal self-injury(NSSI)in young and middle-aged patients with major depressive disorder based on the decision tree algorithm.Methods Young and middle-aged patients with major depressive disorder admitted to the Chongqing Mental Health Center from January 2021 to June 2023 were selected as the research subjects and their clinical data were collected.According to whether NSSI occurred or not,the patients were divided into the NSSI group(n=72)and the non-NSSI group(n=82).Univariate and multivariate logistic regression analyses were used to analyze the risk factors of NSSI in young and middle-aged patients with major depressive disorder.A decision tree predictive model was established based on the results of logistic regression analysis and the Chi-square Automatic Interaction Detection(CHAID)algorithm.The predictive value of the model was evaluated using Receiver Operating Characteristic(ROC)curve.Results The incidence of NSSI in 154 young and middle-aged patients with major depressive disorder was 46.8%.The results of single-factor analysis showed that there were no statistically significant differences(P>0.05)in gender,age,education level,occupation,family composition,only child status,relationship status,fertility status,main place of residence,family economic source,and housing conditions between the two groups.Statistically significant differences(P<0.05)were observed in living conditions,family economic status,duration of depression,SSRS score,PSS-Fa score,and DS score between the two groups.Multi-factor logistic regression analysis showed that lower family economic level,longer duration of depression,SSRS score<20 points,PSS-Fa score<6 points,and DS score≥47.2 points were independent risk factors for NSSI in young and middle-aged patients with major depressive disorder(P<0.05).The constructed decision tree model had five layers and ten nodes,with family economic situation,duration of depression,SSRS score,and PSS-Fa score selected as the nodes

关 键 词：中青年 重度抑郁症 非自杀性自伤行为 影响因素 决策树 

分 类 号：R749.4[医药卫生—神经病学与精神病学]