FDA药物流行病学_真实世界研究(哈佛大学)-2

October2009Hangzhou1Real-lifeStudiesofPharmaceuticalAgents陳建煒,MD,ScD,FISPEHarvardSchoolofPublicHealthi3DrugSafetyOctober17-18,2009October2009Hangzhou2October2009Hangzhou3SometerminologyaboutsafetyTherapeuticandNon-therapeuticuseofmedicalproductsMedicationerrors(adversedrugevents)andbiologiceffectsattherapeuticdose(adversedrugreactions)SafetysignalOctober2009Hangzhou4SafetyofmedicalproductsPatientsafetyAsaresultofhumanerror/systemfailureDrug/Vaccine/DevicesafetyToxicityoccurredwhentheproductwasused‘appropriately’GrayareaInappropriateuse(notonlyoff-labeluse)October2009Hangzhou5ByDr.RobertNelson,inHartzema,Tilson,ChanOctober2009Hangzhou6MethodsandobjectivesInterventionalObservationalstudies(clinicalstudiestrials)BenefitEfficacy(Comparative)PhaseII/IIIEffectivenessRiskPragmatictrialReal-lifeuseofLarge&SimpletrialmedicalproductsOctober2009Hangzhou7ApplicationsofpharmacoepidemiologySafetyassessmentRelativeandabsoluteriskestimatesRiskfactorsforadverseoutcomesUsagepatternsSafetysignaldetectionComparativeeffectivenessDrugutilization(Developmentofmedicalproducts)October2009Hangzhou8ThreeexamplesfortodayandtomorrowStatinsandrhabdomyolysisRiskfactorsforStevensJohnsonSyndromeAprotininincardiacsurgeryandincreasedriskofend-organdamageOctober2009Hangzhou9Why?Pharmacoepidemiology(Therapeutic)riskmanagementOctober2009Hangzhou10JournaloftheNationalCancerInstitute2005;97:872-3“Therealprobleminsafetyisoftennotinthedrugbuthowitisused,”--RaymondWoosley,MD,PhD,CriticalPathInstitute“Inmanyways,afteradrugisreleased,therealtrialbegins,”--PaulSeligman,MD,MPH,OfficeofPharmacoepidemiologyandStatisticalScience,FDAOctober2009Hangzhou11ABriefreview:fromlaboratorytobedside(T1andT2)Fromlaboratorytoapproval(FDA)TranslationalresearchFromapprovaltobedside/community(AgencyforHealthcareResearchandQualityAHRQ)SometimescalleddiffusionoftechnologyTheroleofepidemiologyinthisprocessOctober2009Hangzhou12R&DintheindustryResearchBasic/benchresearchAnimal,cellular,andmolecularmodelsDevelopment‘Bringamoleculetothemarket’EvaluationfromclinicalandbusinessperspectiveOctober2009Hangzhou13ClinicaldevelopmentTheroleofsurrogatemarkersandsurrogateendpointsSolidepidemiologydataarerequiredtosubstantiatethecorrelationbetweensurrogateendpointsandimportantclinicaloutcomesEnd-pointtrialsandOutcometrialsMayormaynotbegoodenoughforsafetyassessmentOctober2009Hangzhou14ClinicaldevelopmentPhaseIIIbandIVNewdosageandformulation(supplementaryNDA)Newindications(supplementaryNDA)LongtermoutcomesEffectiveness(vs.efficacy)SafetyQuality-of-life/Functionalstatus/Cost-effectivenessoutcomesOctober2009Hangzhou15What’swrongwiththecurrentsystemfromasafetyperspective?Limitedgeneralizabilityofanimal,cellular,andmolecularmodelsLimitedutilityofsurrogateendpointsForeffectivenessForsafetyInsufficientknowledgeaboutbackgroundratesandriskfactorsforadverseoutcomesOctober2009Hangzhou16What’swrongwiththecurrentsystemfromasafetyperspective?LimitationsofPhaseIIItrialsComparisongroupDurationPatientpopulationPre-marketingvs.Post-marketingClinicaltrialsvs.ObservationalstudiesThechallengeofcombiningstudies(meta-analysis)MisuseofstatisticsOctober2009Hangzhou17TheScienceofSafety“Weareseeingtheemergenceofascienceofsafety.Thissciencecombinesthegrowingunderstandingofdiseaseanditsoriginsatthemolecularlevel(includingunderstandingofadverseeventsresultingfromtreatment)withnewmethodsofsignaldetection,datamining,andanalysis,enablingresearcherstogeneratehypothesesabout,andconfirmtheexistence,andcausalfactors,ofsafetyproblemsinthepopulationsusingtheproducts.”October2009Hangzhou18DifferentparadigmTwosetsofrelatedmethodology,datasources,interpretationoffindings,andregulatoryimplicationsScienceofefficacyHypothesistestingControlledexperimentsSingleendpoint/outcomeScienceofsafetyRiskidentification/Signaldetection(hypothesisgeneration)PublichealthsurveillanceRiskassessment(hypothesistesting)Experimental(clinicaltrial)andobservational(epidemiology)methodsOctober2009Hangzhou20ScienceofSafetyRiskidentification/signaldetectionHypothesisgenerationRiskassessment/signalconfirmationorrefutationHypothesistestingRiskmitigation/managementOctober2009Hangzhou21DatasourcesforsafetysignalsAnimal/cellularmodelClinicaltrials(pre-andpost-marketing)AdverseeventsLaboratorydata(assurrogatemarkers)Numericalimbalancevs.statisticalsignificanceSpontaneousreports(Pharmacovigilance)October2009Hangzhou22SpontaneousAdverseDrugReactionsReportingSystemsVoluntaryFourelementsAnidentifiablepatientAnadverseeventAsuspecteddrug/vaccine/deviceAnidentifiablereporterSensitivityoverSpecificityOctober2009Hangzhou23SeriousAdverseEventsDeathLifethreateningHospitalization(initialorprolon