精神分裂症急性期患者攻击行为发生情况调查及其Nomograms预测模型的建立  

Investigation on the occurrence of aggressive behavior in patients with acute schizophrenia and establishment of Nomograms prediction model

作　　者：唐红梅 赵林 秦榕蔚 鲁维 Tang Hongmei;Zhao Lin;Qin Rongwei;Lu Wei(Department of Emergency,Sichuan Nanchong Mental Health Center(Nanchong Second People’s Hospital),Nanchong Sichuan 637000,China)

机构地区：[1]四川省南充精神卫生中心,南充市第二人民医院急诊科,四川南充637000

出　　处：《遵义医科大学学报》2023年第2期182-188,共7页Journal of Zunyi Medical University

摘　　要：目的调查精神分裂症急性期患者攻击行为发生情况,并建立其Nomograms预测模型。方法选择2019年1月至2021年3月我院接诊的319例精神分裂症患者作为建模组,另选择2018年1月至12月的精神分裂症104例精神分裂症患者为验证组。收集可能影响精神分裂症急性期患者出现院前攻击行为的因素,以有无院前攻击行为将患者分为2组,比较各影响因素,并行多因素Logistic回归分析。根据多因素分析结果采用R软件构建Nomograms预测模型,并对模型预测效能进行验证。结果本研究纳入的319例精神分裂症患者中共有114例(35.74%)存在院前攻击行为。单因素分析基础上多因素分析结果显示:既往有攻击行为(OR=1.542,95%CI=1.094~2.173)、精神残疾(OR=1.551,95%CI=1.092~2.203)及激越症状(OR=1.570,95%CI=1.067~2.310)为精神分裂症急性期患者院前攻击的危险因素,文化程度为大专及以上(OR=0.662,95%CI=0.466~0.940)、入院前1周抗精神分裂药物服药情况(OR=0.624,95%CI=0.432~0.903)为保护性因素(P<0.05)。根据多因素分析结果以R软件构建Nomograms预测模型,以ROC法评估模型预测价值,模型组AUC为0.770,95%CI为0.714~0.825;验证组AUC为0.767,95%CI为0.711~0.824(P<0.05)。以Bootstrap法对列线图进行内部验证,以原始数据重复抽样1000次,并以验证组进行外部验证,校准曲线结果显示:模型组与验证组预测曲线与标准曲线基本拟合。H-L拟合优度检验结果显示P>0.05。决策曲线分析结果显示:当模型预测概率阈值为0.19~0.80时,临床获益率最高。结论精神分裂症急性期患者院前攻击行为的发生与既往攻击行为、精神残疾、文化程度等因素相关。本研究建立的Nomograms预测模型用于预测精神分裂像患者院前攻击行为具有较高的区分度与准确度,当模型预测概率阈值为0.19~0.80时,临床获益率最高。Objective To investigate the occurrence of aggressive behavior in patients with acute schizophrenia,and to establish the Nomograms prediction model.Methods 319 patients with schizophrenia admitted to our hospital from January 2019 to March 2021 were selected as the modeling group,and 104 patients with schizophrenia from January to December 2018 were selected as the validation group.The factors that may affect the prehospital aggressive behavior of patients with schizophrenia were collected,and the patients were divided into 2 groups with or without prehospital aggressive behavior.The effects of influencing factors on the aggressive behavior were compared,and multi-factor Logistic regression analysis was performed.According to the multivariate analysis results,the Nomograms prediction model was constructed by R software,and the prediction efficiency of the model was verified.Results A total of 114(35.74%)of the 319 patients with schizophrenia included in this study had prehospital aggressive behavior.The results of multi-factor analysis based on univariate analysis showed that previous aggressive behavior(OR=1.542,95%CI=1.094-2.173),mental disability(OR=1.551,95%CI=1.092-2.203)and agitation symptoms(OR=1.570,95%CI=1.067-2.310)were risk factors for prehospital aggression in patients with schizophrenia in the acute phase.Educational level of college degree or above(OR=0.662,95%CI=0.466-0.940)and anti-schizophrenia drug use 1 week before admission(OR=0.624,95%CI=0.432-0.903)were protective factors(P<0.05).According to the multivariate analysis results,the Nomograms prediction model was constructed by R software,and the ROC method was used to evaluate the predictive value of the model.The AUC of the model group was 0.770 with the 95%CI of 0.714-0.825(P<0.05).The nomogram was verified internally by the Bootstrap method,and the original data was repeatedly sampled 1000 times,and the verification group was used for external verification.H-L goodness of fit test results showed(P>0.05).The results of decision curve analysi

关 键 词：精神分裂症 院前攻击行为 多因素分析 Nomograms预测模型 

分 类 号：R749.3[医药卫生—神经病学与精神病学]