临床试验中样本量确定的统计学考虑-专家共识

、、、、、、、、、、、、、、、、、、、、、、、。··CCTSsamplesizedeterminationsamplesizeestimationI、。-。ICHE91998。、、、、。。、、、、、、、、、、、、、effectsize、significantlevel、power、、、。1.1。。2superi-oritytrial、equiv-alencetrial、non-inferioritytrial。、Δ。H0μT=μCH1μT≠μC。P≤α珔XT＞珔XC。μTμC珔XT珔XC。ΔΔ＞0H0μT－μC≤ΔH1μT－μC＞Δ。P≤α。。ΔH0μT－μC≤－ΔμT－μC≥ΔH1－Δ＜μT－μC＜Δ。。Δ1Δ2H0μT－μC≤－Δ1μT－μC≥Δ2H1－Δ1＜μT－μC＜Δ2。P≤α。。ΔH0μT－μC≤－ΔH1μT－μC＞－Δ。P≤α。3、、、、。。2.·727·20158324。II、IIIIV。。。。。、。。、、、、、、。。。3.ＲＲ、HＲ、OＲ、。1。、III、。。2。。。meta。3。。0.8。0.2。0.5。、、、、meta。4.、、、。。Weibull。Iα0.05。α=0.025α=0.0250.05。。Iαα*ααα*α。α0.1。1-ββII。αH0H0。。80%。。。。0.050.0250.05。。。。III23。·827·ChineseJournalofHealthStatisticsAug．2015Vol．32No．41.。θfθTfθT－fθVarT槡～N011VarTT。α1-β。H0fθ=0αH0|T/VarT槡|＞Z1－α/22H1fθ=dε=dVarT槡1－β=P|Nε1|＞Z1－α/2≈PNε1＞Z1－α/2=PN01＞Z1－α/2－ε=Φε－Z1－α/23Z1－βVarT槡=d－Z1－α/2VarT槡VarT=d2/Z1－β+Z1－α/224VarT。。T=珔XE－珔XCδ=μE－μCσ2VarT=σ2nE+σ2nC=r+1r·σ2nC5rnE=r·nC。54δ2Z1－β+Z1－α/22=r+1r·σ2nCnC=r+1r·Z1－β+Z1－α/22σ2δ26σ2。σ2S2tZ．t．。1－β=ΦrnEδ2r+1σ槡2－Z1－α/()27tt1－β=probtt1－α/2nEr+1－2nEr+1－2rnAδ2r+1σ槡()28。accrualperiodATtotaltimeTTfollow-upperi-odTT-AT。ijtijhazardrateλiiηizijATgz=γe－γz/1－γe－γAT0≤z≤ATγ＞0γ＜0γ=0gz=1/AT。xijζijxijζij=mintijTT－zijItij≤TT－zijζ=1ζ=0。T=^λE－^λCδλ=λE－λCVarT=σ2λEnE+σ2λCnC994nE=r·nCnEnCrσ2λEnE+σ2λCnC=δ2λ/Z1－β+Z1－α/22nC=Z1－β+Z1－α/22σ2δ2λσ2λEr+σ2λC()10σ2λσ2λiηiγi=λ2iλiλi+ηi+λiγie－λi+ηiTT1－eλi+ηi－γATλi+ηiλi+ηi－γi1－e－γiAT[]－1i=EC11T=logHＲVarT4num-berofevents。2.、、、。《》20122014“nQuerySAS”199。nQueryAdvisor7.0·927·20158324SAS9.2。1。1ttGreenhouse－Geisser{/ttSatterthwaite’stttWilcoxon/Mann－WhitneyWilcoxon/Mann－WhitneyGreenhouse－Geisser2×2—One－wayANOVATwo－{wayANOVAχ2McNemarχ2χ2{－－－{Pearsonχ2Pearsonχ2Yetesχ2Yetesχ2FisherCMHCMH－－－{—χ2logistic{——————log－ranklog－rank、、、{——————kappaLin{1.。、。。per-protocolset·037·ChineseJournalofHealthStatisticsAug．2015Vol．32No．4。20%meta。。、、。。2.。。1。2。3。9。。。Q1.nQueryAdvisor+nTerimPASSDSTPLANG*PowerPC-SizePSSASPowerandSampleSizeapplicationPSSStataＲ。nQueryPASS。。nQueryAdvisor+nTerimStatisticalSolu-tionsnQueryAdvisor7nTerim。3.0Windows。FDA、、、5049。。《》9nQueryAdvisor7.0SAS9.2。PASSNCSSWin-dows13。nQuery230。DSTPLANWindowsFortran。t、、、2×N。G*PowerWindowsMacOSX。t、One-wayANOVA、、。。PC-SizeWindowsDOS。t、、、。。PSWindowst、、Fisher、McNemar、。SAS/PowerandSampleSizeapplicationPSSWindowsSAS。SASt、、、、。Stata/ＲStataＲ。。733·137·20158324、、。、、、、、、、。1..20101915221.2.———.201027254-59.3..2010342169-170.4..2008152128-129.5.20052009.201081962-64.6.2007-05-09.7.45-59.201011831.8.2007-20113030.20122771032-1035.9.1993-2011.20121297-99.731nQueryPASS。Q2.。II。。Q3.1-ββII。ββ。。Q4.。III。。。CCTS。Q5.。。1ICHE10ChoiceofControlGroupandＲelatedIssuesinClinicalTrials．USDepartmentofHealthandHumanServicesFoodandDrugAdministration2001．2ICHE9StatisticalPrinciplesforClinicalTrials．USDepartmentofHealthandHumanServicesFoodandDrugAdministration1998．3《》2003914CCTS．．2012293445-4505JuliousSA．SampleSizesforClinicalTrials．BocaＲatonChapman＆Hall/CＲC2010．6MachinDCampbellMJTanSBetal．SampleSizeTablesforClini-calStudies．UKWiley-Blackwell3rdedit2009．7ＲyanTP．SampleSizeDeterminationandPower．USAJohnWiley＆SonsInc2013．8ChowSC．SampleSizeCalculationsinClinicalＲesearch．USAChapman＆Hall/CＲC2ndEdit2008．9．nQuerySAS19．2012291-2014312．10ElashoffJD．nQueryAdvisorUser'sGuide．IrelandStatisticalSolu-tionsLtd．2007．11SASInstituteInc．GettingStartedwiththeSASPowerandSampleSizeApplication．NorthCarolinaSASInstituteInc．2004．12CrispACurtisP．Samplesizeestimationfornon-inferioritytrialsoftime-to-eventdata．Pharmaceuticalstatistics20087236-244．·337·ChineseJournalofHealthStatisticsAug．2015Vol．32No．4