摘要
目的:采用Logistic回归构建精神分裂症复发风险预测模型。方法:采用一般情况调查表、中医体质量表、精神分裂症复发先兆量表(ESS)、领悟社会支持量表(PSSS)和服药依从性评定量表(MARS,MMAS-8)调查湖南省1 013例稳定期精神分裂症患者,根据精神分裂症复发先兆量表,以21.5分为临界值,分为复发风险组和无复发风险组。运用Logistic回归分析精神分裂症复发的风险因素,受试者工作特征曲线(ROC)检测模型的预测效能。结果:对精神分裂症复发相关危险因素的调查分析结果显示,居住在农村(OR=0.702,P<0.05)、文化程度低(OR=0.785,P<0.05)、出院后未就业(OR=0.716,P<0.05)、社会支持状态低(OR=0.726,P<0.05)、中医体质属于气虚质(OR=0.042,P<0.05)、阴虚质(OR=0.175,P<0.05)、湿热质(OR=0.127,P<0.05)及服药依从性低(OR=0.467,P<0.05)均为影响精神分裂症复发的独立危险因素。影响精神分裂症患者复发的Logistic回归模型为:Logit(p)=5.405-0.354X1(居住地)-0.241X2(文化程度)-0.334X4(出院后是否就业)-0.320X5(领悟社会支持)-3.176X(6(1))(气虚质)-1.745X(6(3))(阴虚质)-2.062X(6(5))(湿热质)-0.761X8(服药行为)。该模型准确率为75.9%,通过HosmerLemeshow检验,Hosmer-Lemeshow χ^(2)=3.687,P=0.884>0.1,说明模型与观测值拟合度较好。Logistic回归模型预测精神分裂症复发的AUC为0.814,95%CI为(0.778~0.841)。结论:基于Logistic回归建立精神分裂症复发风险预测模型具有较好预测能力,临床可采用Logistic回归模型分析各变量与精神分裂症复发的相关性并予以早期干预。
Objective:To establish the construction of risk prediction model for schizophrenia relapse with Logistic regression.Methods:A survey was conducted on 1013 stable period schizophrenia patients in Hunan Province using a general data questionnaire,traditional Chinese medicine constitution Scale,the Early Signs Scale(ESS),Perceived Social SuPPort Scale(PSSS),and Medication Adherence Rating Scale(MARS,MMAS-8).According to the ESS,a score of 21.5 was used as the critical value,and they were divided into recurrence risk group and non recurrence risk group.Logistic regression analysis was used to analyze the risk factors for relapse of schizophrenia,and the predictive performance of the receiver operating characteristic curve(ROC)test model was evaluated.Results:The investigation and analysis of risk factors related to the recurrence of schizophrenia showed that living in rural areas(OR=0.702,P<0.05),having a low education level(OR=0.785,P<0.05),not being employed after discharge(OR=0.716,P<0.05),having a low social support status(OR=0.726,P<0.05),having a traditional Chinese medicine constitution of Qi deficiency(OR=0.042,P<0.05),Yin deficiency(OR=0.175,P<0.05),damp heat(OR=0.127,P<0.05),and low medication adherence(OR=0.467,P<0.05)were independent risk factors for the recurrence of schizophrenia.The logistic regression model that affects the recurrence of schizophrenia patients was:Logit(p)=5.405-0.354X1(place of residence)-0.241X2(education level)-0.334X4(whether employed after discharge)-0.320X5(understanding social support)-3.176X6(1)(qi deficiency)-1.745X6(3)(yin deficiency)-2.062X6(5)(damp heat)-0.761X8(medication behavior).The accuracy of this model was 75.9%,and it passed the Hosmer Limeshow testχ^(2)=3.687,P=0.884>0.1,indicating a good fit between the model and the observed values.The logistic regression model predicted AUC of 0.814 for relapse of schizophrenia,with a 95%CI of(0.778~0.841).Conclusion:The establishment of a risk prediction model for schizophrenia relapse based on logistic regression has good pr
作者
王伟月
蒋小剑
向婷婷
刘娟
葛盈琼
朱正敏
邹书亭
谭碧璨
WANG Weiyue;JIANG Xiaojian;XIANG Tingting;LIU Juan;GE Yingqiong;ZHU Zhengmin;ZOU Shuting;TAN Bican(Hunan University of Chinese Medicine,Changsha Hunan 410208,China;Hunan Province Maternal and Child Health Care Hospital,Changsha Hunan 410028,China;Ningxia Medical University,Yinchuan Ningxia750004,China)
出处
《中医药导报》
2023年第4期92-97,共6页
Guiding Journal of Traditional Chinese Medicine and Pharmacy
基金
湖南省创新型省份建设专项经费资助项目(2021SK2009)
宁夏回族自治区重点研发计划(对外科技合作专项)项目(2019BFG02023)
湖南省普通高等学校教学改革研究项目(HNJG-2022-0140)。
