资源描述：
《lecture8_statistical analysis of stability data》由会员上传分享，免费在线阅读，更多相关内容在教育资源-天天文库。
1、LectureEightStatisticalAnalysisofStabilityDataOUTLINEIntroductionSingleBatchesMultipleBatchesPreliminaryTestforBatchesFixedEffectsModelMinimumApproachMultipleComparisonProcedureforPoolingBatchesRandomEffectsModelChowandShao'smethodRandomRegressionCoefficientHo,LiuandChow's(HLC)MethodOUTLINENumer
2、icalExamplesSimulationResultsComparisonofMethodsOtherIssuesI.IntroductionStabilitydatafromlong-termstudiesunderambientconditions(nonaccelerateddata).ExpirationDatingPeriod(Shelf-life)FDAGuideline(1987)p.29"toestablish,withahighdegreeofconfidence,anexpirationdatingperiodduringwhichtheaveragedrugp
3、roductcharacteristic(i.e.,strength)ofthebatchwillremainwithinspecifications.Thisexpirationdatingperiodshouldbeappliedtoallfuturebatches..."p.30"Also,percentoflabelclaim,notpercentofinitialaverage-value,isthevariableofinterest."I.IntroductionPDAGuideline(1987)p.31"Anacceptableapproachfordrugchara
4、cteristicsthatareexpectedtodecreasewithtimeistodeterminethetimeatwhichthe95%one-sidedlowerconfidencelimit...formeandegradationcurveintersectstheacceptablelowerspecificationlimit."p.32"...,wemaybe95%confident&thattheaveragedrugproductcharacteristic(i.e.,strength)ofthedosageunitsinthebatchiswithin
5、specificationsuptotheendoftheexpirationdatingperiod."II.SingleBatchOnlyconsiderthecasewherethedrugproduct characteristicdecreaseslinearlywithtime.Model:(2.1):jthresponseofassayattimeXj,:Intercept(batcheffect),β:Slope(degradationrate),Xj:timeatwhichYjisobserved,εj:randomerror~N(0,2).SAS:PERCENT
6、=TIMELSE:theMSEfromthemodel(1-α)100%lowerC.L.forthemeandegradationcurve(i.e.,α+βX)atX:η:theacceptablelowerspecificationlimit(e.g.,90%)TworootsXLandXUare(l-2α)100%C.I.for(η-α)/βConditions:(a)(η-a)/SE(a)<-t(α,n-2)(b)b/SE(b)<-t(α,n-2)(1)If(a)and(b)hold,(l-2α)100%C.I.for(η-α)/βisinclusive.Estimateds
7、helf-lifeisXL(Kohberger,1988).(2)If(b)doesnotholdandb<0,(l-2α)100%C.I.for(η-α)/βis(XL,).Estimatedshelf-lifeisXL(Kohberger,1988)Esterling(1969)Construct(l-2α)100%C.I.forXforwhichthepthupperquantileofthedistributionofYgivenXis
