住院精神分裂症患者冲动行为的风险预测模型的构建与应用  

Construction and application of risk prediction model for impulsive behavior of inpatients with schizophrenia

作　　者：汪敏[1] 朱小英[1] 李玲丽[1] WANG Min;ZHU Xiao-ying;LI Ling-li(Hangzhou Seventh People's Hospital,Hangzhou 310013,China)

机构地区：[1]杭州市第七人民医院精神四科,310013

出　　处：《临床精神医学杂志》2023年第2期133-136,共4页Journal of Clinical Psychiatry

摘　　要：目的:探究住院精神分裂症患者冲动行为的风险预测模型的构建与应用情况。方法:选取334例精神分裂症患者,根据出院后外显攻击行为量表(MOAS)评分分为冲动行为组(n=116)和非冲动行为组(n=218),应用Logistic回归分析冲动行为的相关影响因素,根据分析结果建立风险预测模型,选取116例患者进行模型验证,并分析其临床预测价值。结果:冲动行为组在性别、饮酒史、不良生活习惯、既往冲动行为史、被害妄想、易激惹、幻听、关系妄想、敌对情绪、兴奋方面比例与非冲动行为组差异有统计学意义(P<0.05或P<0.01)。经Logistic多因素分析后,以性别、不良生活习惯史、既往冲动行为史、易激惹为精神分裂症患者发生冲动行为的独立危险因素。风险预测模型预测ROC曲线下面积为0.769(95%CI:0.712~0.826,P<0.001),约登指数最大值为0.462,敏感度为0.595,特异度为0.867。结论:以性别、不良生活习惯史、既往冲动行为史、易激惹为危险因素建立风险预测模型能够有效预测住院精神分裂症患者冲动行为的发生,且具有较高的特异度和灵敏度,对临床干预具有一定指导意义。Objective:To explore the construction and application of risk prediction model of impulsive behavior in hospitalized schizophrenic patients.Method:A total of 334 schizophrenics were selected and divided into impulsive behavior group(n=116)and non-impulsive behavior group(n=218)according to Modified Overt Aggression Scales(MOAS)score after discharge.Logistic regression was applied to analyze the relevant influencing factors of impulsive behavior,and a risk prediction model was established according to the analysis results.116 patients were selected for model validation,and its clinical prediction value was analyzed.Results:The proportions of impulsive behavior group in gender,drinking history,bad living habits,history of previous impulsive behavior,delusion of victimization,irritability,auditory hallucinations,relationship delusion,hostility and excitement were significantly different from those of the non-impulsive behavior group(P<0.05 or P<0.01).After Logistic multivariate analysis,gender,history of bad living habits,history of previous impulsive behavior and irritability were independent risk factors for impulsive behavior in patients with schizophrenia.The area under ROC curve predicted by risk prediction model was 0.769(95%CI:0.712-0.826,P<0.001),the maximum value of Youden index was 0.462,the sensitivity was 0.595 and the specificity was 0.867.Conclusion:The establishment of risk prediction model with gender,history of bad living habits,history of previous impulsive behavior and irritability as risk factors can effectively predict the occurrence of impulsive behavior in hospitalized schizophrenic patients,and has high specificity and sensitivity,which has important guiding significance for clinical intervention.

关 键 词：精神分裂症 风险预测模型 冲动行为 预测价值 

分 类 号：R749.3[医药卫生—神经病学与精神病学]