您好,欢迎访问三七文档
当前位置:首页 > 医学/心理学 > 药学 > 在多中心临床试验中对中心校正后疗效的统计分析方法-何清波
1.(200025)2.3.4.何清波1 苏炳华1 姚 晨2 陈 峰3 孙 高4 ,。,:、、。,WilcoxonRidit。,,,,2×2。,,,,〔1,2〕。,。,CMH(Cochran-MantalHaenszel)logistic。SASFREQLOGISTIC。,。:,100,AB50。: 1A101714950B5112014502A141414850B5131121503A216131050B714191050 ,。:dataa; doCenter=1to3: dogroup=1to2; doresp=1to4; ifresp=2thenresp2=1;elseresp2=2; inputf @@; doi=1tof; output; end; end;end;end;Cards;101714951120141414148513112121613107141910;run;,SASa,300,。acenter,group(A1,B2),resp(1、2、34、、)。resp2=1,resp2=2,。:,,,、、。,,。1.CMH,SAS:procfreqdata=a;tablescenter*group*resp/nopercentnocolchisqscores=rankcmh2;run;tablescenter*group*resp,RESP,GROUP,CENTER·252· 20008174,,。scores=rankcmh2,Wilcoxon(,Kruskal-Wallis),〔3,4〕。,:TABLE1OFGROUPBYRESPCONTROLLINGFORCENTER=1GROUPRESPFrequencyRowPct12341101714920.0034.0028.0018.002511201410.0022.0040.0028.00Total15283423Total5050100STATISTICSFORTABLE1OFGROUPBYRESPCONTROLLINGFORCENTER=1StatisticDFValueProbChi-Square35.0980.165LikelihoodRatioChi-Square35.1550.161MHChi-Square(RankScores)14.4950.034PhiCoefficient0.226ContingencyCoefficient0.220Cramer′sV0.226SampleSize=100TABLE2OFGROUPBYRESPCONTROLLINGFORCENTER=2GROUPRESPFrequencyRowPct12341141414828.0028.0028.0016.002513112110.0026.0022.0042.00Total19272529Total5050100STATISTICSFORTABLE2OFGROUPBYRESPCONTROLLINGFORCENTER=2StatisticDFValueProbChi-Square310.4880.015LikelihoodRatioChi-Square310.8770.012MHChi-Square(RankScores)18.5160.004PhiCoefficient0.324ContingencyCoefficient0.308Cramer′sV0.324SampleSize=100TABLE3OFGROUPBYRESPCONTROLLINGFORCENTER=3GROUPRESPFrequencyRowPct12341216131042.0012.0026.0020.002714191014.0028.0038.0020.00Total28203220Total5050100STATISTICSFORTABLE3OFGROUPBYRESPCONTROLLINGFORCENTER=3StatisticDFValueProbChi-Square311.3250.010LikelihoodRatioChi-Square311.7480.008MHChi-Square(RankScores)13.0950.079PhiCoefficient0.337ContingencyCoefficient0.319Cramer′sV0.337SampleSize=100SUMMARYSTATISTICSFORGROUPBYRESPCONTROLLINGFORCENTERCochran-Mantel-HaenszelStatistics(BasedonRankScores)StatisticAlternativeHypothesisDFValueProb1NonzeroCorrelation115.4100.0012RowMeanScoresDiffer115.4100.001TotalSampleSize=300,。,。MHChi-Square(RankScores)Wilcoxon。1、234.495、8.5163.095,1,P0.034、0.0040.079。,B;12,3。,Wilcoxon,QCMH=15.410,P=0.001。,BA。,TABLESrespresp2,scores=rank。(Chi-Square)。2.logistic,SAS:datab;seta;ifcenter=1thencen1=1;elsecen1=0;ifcenter=2thencen2=1;elsecen2=0;·253· ChineseJournalofHealthStatistics,August2000,Vol.17,No.4proclogisticdata=b; modelresp=groupcen1cen2;run;,。。,,,datab。,:CEN1CEN2。1CEN1=1CEN2=0;2CEN1=0CEN2=1;3CEN1CEN20。modelgroup、cen1cen2。model,,resp。:TheLOGISTICProcedureDataSet:WORK.BResponseVariable:RESPResponseLevels:4NumberofObservations:300LinkFunction:LogitResponseProfileOrderedValueRESPCount1162227533914472ScoreTestfortheProportionalOddsAssumptionChi-Square=11.6620with6DF(P=0.0699)ModelFittingInformationandTestingGlobalNullHypothe-sisBETA=0InterceptInterceptandCriterionOnlyCovariatesChi-SquareforCovariatesAIC832.065819.698·SC843.177841.920·-2LOGL826.065807.69818.368with3DF(P=0.0004)Score··17.302with3DF(P=0.0006)AnalysisofMaximumLikelihoodEstimatesParameterStandardWaldPrVariableDFEstimateErrorChi-SquareChi-SquareINTERCP110.14600.36410.16080.6885INTERCP211.37000.370613.66800.0002INTERCP312.74340.395948.01430.0001GROUP1-0.84470.212215.84680.0001CEN11-0.38600.25542.28400.1307CEN21-0.40090.25552.46200.1166 ,、、、。,ResponseProfile、。,,。(ScoreTestfortheProportionalOddsAssumption)。,〔5〕。,χ2=11.6620,df=6,P=0.06990.05,。β=0(ModelFittingInformationandTestingGlobalNullHypothesisBETA=0)。(Chi-SquareforCovariates)P0.00040.0006,,。。groupWald15.8468,P=0.0001,。group(-0.8447),group(group=2,B),resp();BA。CMH。cen1cen2P0.13070.1166,。,。,MODELrespresp2。,SAS。。 1.,.———.:,1997,44.2.ICHTopicE9.StatisticalPrinciplesforClinicalTrials.1997,622.3.,.SASSAS/STAT.:,1997,70,394.4.SASInstituteInc.SAS/STATUser'sGuide.Version6.FourthEdition.Northcarolina,USA.SASInstituteInc.1994,128,885-887.5..SAS.:,1996,174-177.·254· 20008174
本文标题:在多中心临床试验中对中心校正后疗效的统计分析方法-何清波
链接地址:https://www.777doc.com/doc-4472481 .html