欢迎来到天天文库
浏览记录
ID:58296500
大小:305.12 KB
页数:4页
时间:2020-05-02
《利用两水平logistic模型探讨麻疹监测系统病例分类.pdf》由会员上传分享,免费在线阅读,更多相关内容在行业资料-天天文库。
1、中华流行病学杂志2014年1月第35卷第1期ChinJEpidemiol,January2014,Vo1.35,No.1·57··监『J·利用两水平logistic模型探讨麻疹监测系统病例分类黎翰东敖睿彭琳李迪张菊英【摘要】目的建立判别麻疹临床症状的logistic回归模型,比较麻疹与其他发热出疹性疾病临床症状的差异。方法收集2011年3月至2012年2月四川省4个县级以上医院报告的所有麻疹、风疹、猩红热、幼儿急疹等病例,以实验室确诊麻疹为反应变量,各临床症状为自变量建立logistic回归方程,并绘制ROC曲线,选择最佳截断点。结果共收集551名病例
2、,麻疹和风疹临床诊断与实验室诊断一致性均较差,Kappa值=0.349;根据两水平logistic回归分析,与麻疹确诊病例有关的症状有咳嗽(OR=5.75)、结膜炎(OR=3.O0)、柯氏斑(OR=7.52)、淋巴结肿大(OR=0.07)、热退疹出(0=0.07);回归方程ROC曲线的曲线下面积为0.97,最佳截断点为0.249。结论通过临床症状建立了预测效果较好的logistic回归模型,但应开展更大范围的调查和实验室检测,明确与麻疹相关的主要发热出疹性疾病类型及构成。【关键词】麻疹;监测;两水平logistic回归;ROC曲线Caseclassif
3、icationinmeaslessurveillancesysternundertheTwo.1evelLogi【sticModelLiHandong,ADRui,Lin。,LiDi,ZhangJuying.1DepartmentofEpidemiologyandBiostatistics,tSchoolofPublicHeahh,SichuanUniversity,Chengdu610041,China;2SichuanProvincialCenterrDiseaseControlandPreventionCorrespondingauthor:Zh
4、angJuying,Email:juying109@163.corn【Abstract】0bjectiveTostudytheprevalenceratesofRashandFebrileIllnesses(RFIs)includingmeasles,rubella,scarletfever,exanthemasubitumandthediferencesamongmeaslesandotherRFIstotentativelyformulatethelogisticregressionmodelthroughclinicalmanifestation
5、.MethodsAllthesuspectedcasesofmeasles,rubella,scarletfever,exanthemasubitumreportedbythecounty/prefectureleverhospitalsatfourcountieswerecollectedduringMarch2011toFebruary2012.Whensettinglaboratoryconfirmedmeaslesasdependentvariableandexistedsymptomsasindependentvariable.alogist
6、icregressionmodelwasformulatedandoptimaloperationalpoint(OOP)chosen.accordingtotheROCcurve.ResultsAtotalnumberof551caseswerecollectedbuttheconsistencyofmeaslesdiagnosisbetweenclinicalandlaboratorywasnotsatisfied.withKappavalue=0.349.sametothediagnosisofrubella.Asfortheresultfrom
7、thetwo.1everlogisticregressionmode1.syml:Itomsthatrelatedtotheconfirmationofmeasleswouldincludecough(OR=5.75).conjunctivitis(OR=3.00),Koplikspot(OR=7.52),lymphadenectasis(OR=0.07),rashafterfever(OR=0.07).TheareaunderROCcurvewas0.97andtheoptimaloperationalpointwas0.249.Conclusion
8、Alogisticregressionmodelwasformulatedusingthecl
此文档下载收益归作者所有