FDA药物流行病学_真实世界研究(哈佛大学)-2.pptx

《FDA药物流行病学_真实世界研究(哈佛大学)-2.pptx》由会员分享，可在线阅读，更多相关《FDA药物流行病学_真实世界研究(哈佛大学)-2.pptx（100页珍藏版）》请在淘文阁 - 分享文档赚钱的网站上搜索。

1、2021/9/13Hangzhou1 Real-life Studies of Pharmaceutical Agents陳建煒,MD,ScD,FISPEHarvardSchoolofPublicHealthi3DrugSafetyOctober17-18,20092021/9/13Hangzhou22021/9/13Hangzhou3SometerminologyaboutsafetylTherapeuticandNon-therapeuticuseofmedicalproductslMedicationerrors(adversedrugevents)andbiologiceffectsa

2、ttherapeuticdose(adversedrugreactions)lSafetysignal2021/9/13Hangzhou4Safety of medical productslPatientsafetylAsaresultofhumanerror/systemfailurelDrug/Vaccine/DevicesafetylToxicityoccurredwhentheproductwasusedappropriatelylGrayarealInappropriateuse(notonlyoff-labeluse)2021/9/13Hangzhou5ByDr.RobertNe

3、lson,inHartzema,Tilson,Chan2021/9/13Hangzhou6Methods and objectivesInterventionalObservationalstudies(clinicalstudiestrials)BenefitEfficacy(Comparative)PhaseII/IIIEffectivenessRiskPragmatictrialReal-lifeuseofLarge&Simpletrialmedicalproducts2021/9/13Hangzhou7Applications of pharmacoepidemiologylSafet

4、yassessmentlRelativeandabsoluteriskestimateslRiskfactorsforadverseoutcomeslUsagepatternslSafetysignaldetectionlComparativeeffectivenesslDrugutilizationl(Developmentofmedicalproducts)2021/9/13Hangzhou8Three examples for today and tomorrowlStatinsandrhabdomyolysislRiskfactorsforStevensJohnsonSyndromel

5、Aprotininincardiacsurgeryandincreasedriskofend-organdamage2021/9/13Hangzhou9lWhy?lPharmacoepidemiologyl(Therapeutic)riskmanagement2021/9/13Hangzhou10JournaloftheNationalCancerInstitute2005;97:872-3l“Therealprobleminsafetyisoftennotinthedrugbuthowitisused,”-RaymondWoosley,MD,PhD,CriticalPathInstitute

6、l“Inmanyways,afteradrugisreleased,therealtrialbegins,”-PaulSeligman,MD,MPH,OfficeofPharmacoepidemiologyandStatisticalScience,FDA2021/9/13Hangzhou11A Brief review:from laboratory to bedside(T1 and T2)lFromlaboratorytoapproval(FDA)lTranslationalresearchlhttp:/www.fda.gov/oc/initiatives/criticalpath/lh

7、ttp:/www.c-path.org/lFromapprovaltobedside/community(AgencyforHealthcareResearchandQualityAHRQ)lSometimescalleddiffusionoftechnologylTheroleofepidemiologyinthisprocess2021/9/13Hangzhou12R&D in the industrylResearchlBasic/benchresearchlAnimal,cellular,andmolecularmodelslDevelopmentlBringamoleculetoth

8、emarketlEvaluationfromclinicalandbusinessperspective2021/9/13Hangzhou13Clinical developmentlTheroleofsurrogatemarkersandsurrogateendpointslSolidepidemiologydataarerequiredtosubstantiatethecorrelationbetweensurrogateendpointsandimportantclinicaloutcomeslEnd-pointtrialsandOutcometrialslMayormaynotbego

9、odenoughforsafetyassessment2021/9/13Hangzhou14Clinical developmentlPhaseIIIbandIVlNewdosageandformulation(supplementaryNDA)lNewindications(supplementaryNDA)lLongtermoutcomeslEffectiveness(vs.efficacy)lSafetylQuality-of-life/Functionalstatus/Cost-effectivenessoutcomes2021/9/13Hangzhou15Whats wrong wi

10、th the current system from a safety perspective?lLimitedgeneralizabilityofanimal,cellular,andmolecularmodelslhttp:/content.nejm.org/cgi/content/abstract/333/17/1099lLimitedutilityofsurrogateendpointslForeffectivenesslForsafetylInsufficientknowledgeaboutbackgroundratesandriskfactorsforadverseoutcomes

11、2021/9/13Hangzhou16Whats wrong with the current system from a safety perspective?lLimitationsofPhaseIIItrialslComparisongrouplDurationlPatientpopulationlPre-marketingvs.Post-marketinglClinicaltrialsvs.ObservationalstudieslThechallengeofcombiningstudies(meta-analysis)lMisuseofstatistics2021/9/13Hangz

12、hou17The Science of Safety“Weareseeingtheemergenceofascience of safety.Thissciencecombinesthegrowingunderstandingofdiseaseanditsoriginsatthemolecularlevel(includingunderstandingofadverseeventsresultingfromtreatment)withnewmethodsofsignaldetection,datamining,andanalysis,enablingresearcherstogenerateh

13、ypothesesabout,andconfirmtheexistence,andcausalfactors,ofsafetyproblemsinthepopulationsusingtheproducts.”2021/9/13Hangzhou18http:/www.fda.gov/Safety/FDAsSentinelInitiative/ucm089474.htm2021/9/13Hangzhou19Science of Efficacy and Science of SafetylDifferentparadigmlTwosetsofrelatedmethodology,datasour

14、ces,interpretationoffindings,andregulatoryimplicationslScienceofefficacylHypothesistestinglControlledexperimentslSingleendpoint/outcomelScienceofsafetylRiskidentification/Signaldetection(hypothesisgeneration)lPublichealthsurveillancelRiskassessment(hypothesistesting)lExperimental(clinicaltrial)andob

15、servational(epidemiology)methods2021/9/13Hangzhou20Science of SafetylRiskidentification/signaldetectionlHypothesisgenerationlRiskassessment/signalconfirmationorrefutationlHypothesistestinglRiskmitigation/management2021/9/13Hangzhou21Data sources for safety signalslAnimal/cellularmodellClinicaltrials

16、(pre-andpost-marketing)lAdverseeventslLaboratorydata(assurrogatemarkers)lNumericalimbalancevs.statisticalsignificancelSpontaneousreports(Pharmacovigilance)2021/9/13Hangzhou22Spontaneous Adverse Drug Reactions Reporting SystemslVoluntarylFourelementslAnidentifiablepatientlAnadverseeventlAsuspecteddru

17、g/vaccine/devicelAnidentifiablereporterlSensitivityoverSpecificity2021/9/13Hangzhou23Serious Adverse Eventslhttp:/www.fda.gov/medwatch/report/DESK/advevnt.htmlDeathlLifethreateninglHospitalization(initialorprolonged)lDisabilitylCongenitalanomalylRequiresInterventiontoPreventPermanentImpairmentorDama

18、ge2021/9/13Hangzhou24Strengths of the spontaneous reports systemlModeratecostslAcaseorcaseseriescantellacompellingstory(temporalsequence,de-challenge&re-challenge)lMinimaldelaylWorkswellwithpublichealth-orientedclinicians2021/9/13Hangzhou25LimitationslAnecdotalinformationl“Thepluralofanecdoteisnotda

19、ta”l(Usually)poordataqualitylCannotconclusivelyestablishcausalitylReportingratesaffectedbymanysecularfactorslHypothesisgeneration,notconfirmatory2021/9/13Hangzhou26Hypothetical case#1Sudden deathlYear1988,62year-oldmale,noCVriskfactor,survivedacuteMIlFrequentprematureventricularcomplexes(PVCs)afterM

20、IlPatientwasgivenencainide,aclassIcanti-arrhythmicagent,tosuppressthePVCsl6monthslater,patientdiedsuddenlylBackgroundrate:one-yearmortality10%2021/9/13Hangzhou27Would anyone suspect that death was precipitated by the drug?lWhatiftherewasnoCASTtrial?lLearyWE.WarningIssuedon2HeartDrugsAfterDeathsofPat

21、ientsinTest.New York Times,April26,1989,page1A.lCardiacArrhythmiaSuppressionTrial(CAST)Investigators.Preliminaryreport:N Engl J Med1989;321:406-12.(Aug10issue)2021/9/13Hangzhou28Hypothetical case#2Liver failurelYear1999,53year-oldmale,BMI31,socialdrinker,onglyburide,troglitazone,atenolol,simvastatin

22、,anddiclofenaclPatientdevelopedacuteliverfailureanddiedwhilewaitingforlivertransplantationlLiverfailuremaybecausedbyadruglbutwhichdrugorcombinationofdrugs?2021/9/13Hangzhou29What is the difference between case#1 and case#2?lSuspicionandreportingofAdverseDrugReactionslSeverityofeventlPriorknowledgelA

23、lternativeexplanationslBackgroundrateoftheadverseeventlRarelNotsorare2021/9/13Hangzhou30Rare and not-so-rare adverse eventslRarelRhabdomyolysis,DILI,Stevens-JohnsonSyndrome,agranulocytosis,Churg-StraussSyndrome,lNotsorare(especiallyamongcertainpopulations)lMyocardialinfarctionlTuberculosisamongRApat

24、ientslChronicactivehepatitisBexacerbationlSuicideamongadolescentsl,2021/9/13Hangzhou31Hypothetical case#3Acute myocardial infarctionlYear2000,49year-oldmale,smoker,BMI33,mildhypertensionandLDLelevationlLeftkneeosteoarthritislGivenrofecoxib25mg/dayfortwoweeks,thenthedoseincreasedto50mg/dayl9monthslat

25、er,hadnon-fatalmyocardialinfarctionlBackgroundrateofAMI:severalcasesper1,000person-years2021/9/13Hangzhou32An interesting questionlCouldtheexistingsystemidentifyCVsafetysignalsassociatedwithrofecoxibiftherewasnoVIGORtrial?lPUBMEDsearchonOct24,2005,publishedcasereportsofrofecoxib:hepatitis,interstiti

26、alnephritis,colitis,angioedema,anaphylacticshock,gynecomastia,acuterenalfailure,delirium,Stevens-Johnsonsyndrome,congestiveheartfailure,transientvisualimpairment,aquagenicwrinklingofthepalms,2021/9/13Hangzhou33Epidemiology,a general outlinelMeasuresofDiseaseOccurrencelIncidenceandPrevalencelCumulati

27、veincidenceandincidenceratelConceptofdynamicpopulationandpersontimelMeasuresofAssociationlRelativerisklRateratiolOddsratio2021/9/13Hangzhou34Epidemiology,a general outlinelDescriptiveEpidemiologylCasereportslCaseserieslFrequencyofeventslStudiesprovidingsomeevidence,butnotdefinitivelCross-sectionalst

28、udieslEcologicalstudies2021/9/13Hangzhou35Epidemiology,a general outlinelObservationalstudies,selectedmethodslCohortstudieslCase-controlstudieslCase-crossoverstudieslInterventionalstudieslClinicaltrialsforindividualsubjectslClusteredrandomizedtrials2021/9/13Hangzhou36Epidemiology,a general outlinelC

29、ombininginformationfrommultiplestudieslMeta-analysislDatapooling2021/9/13Hangzhou37Systematic and random errorslValidityconsiderationinepidemiologylConfoundinglBiaslInformationbiaslSelectionbiaslAssessmentofrandomerrorinepidemiology:statistics2021/9/13Hangzhou38Commonly used terms in epidemiologyCon

30、founder(s)ExposureOutcome2021/9/13Hangzhou39Exposures in pharmacoepidemiologylDrugslVaccineslMedicaldeviceslProcedureslProgramsle.g.screeningprogramslCodingsystemsforeachofthem2021/9/13Hangzhou40Ascertainment of OutcomeslOperationaldefinitionslSensitivityandspecificitylObjectivecriterialE.g.hemorrha

31、gicstrokelBlindedadjudicationifpossible2021/9/13Hangzhou41Performance of ICD-9 codes in the study of rhabdomyolysis2021/9/13Hangzhou42Andradeet al.Healthplanadministrativedatabasescanefficientlyidentifyseriousmyopathyandrhabdomyolysis.JClinEpi2005;58:171-42021/9/13Hangzhou43Cohort studies in pharmac

32、oepidemiologylSourcepopulationlAscertainmentofexposurelComparisongroups/cohortslConfounding?lAscertainmentofoutcomeslInformationBias?lControlforconfoundinglEvaluationofeffectmodificationlStrengthsandlimitations2021/9/13Hangzhou44Exposure and OutcomelExposuresofinterest/ComparisongroupslHead-to-headc

33、omparisonlDrugvs.DiseasecomparisonlMultiplecomparisongroupslOutcomesofinterestlClinicallyrelevant2021/9/13Hangzhou45Arch Intern Med 2002;162:936-42.lThreestudygroupslAlendronateusers(n=6,432)lWomenwithosteoporoticfracture(n=9,776)lAge-gendermatchedunexposed(n=33,176)lComparingalendronateusersandunex

34、posedlAdjustedrateratio=1.8,95%confidenceinterval0.83.9lComparingalendronateusersandwomenwithfracturelAdjustedrateratio=1.1,95%confidenceinterval,0.6-2.22021/9/13Hangzhou46Defining cohortslMatchinglIndividually-matchlFrequency-matchlMatchingincohortstudyvs.matchingincase-controlstudylRestriction/Sel

35、ectioncriterialNew-userdesignlIncidentusersvs.Prevalentusers2021/9/13Hangzhou47Bias and confoundinglConfoundinglChanneling,Confounding-by-Indication,andConfounding-by-SeveritylHowtomeasureindicationandseverity?lOtherknownriskfactorsfortheoutcomeofinterest(age,gender,smoking)lBiaslDetectionbiasasafor

36、mofinformationbias2021/9/13Hangzhou48Analytic approacheslAtriskperson-timelImmortalperson-timelCumulativeincidence,incidence,andcorrelatedoutcomeslCount,person-time,andtime-to-eventlControlforconfoundinglMultipleregression,propensityscore,instrumentalvariables,andotherslAccountforunmeasuredconfounde

37、rslNestedcase-controlstudy2021/9/13Hangzhou49Stratified analysislControlforconfoundinglEffectmodificationlSpecificsubgroup(s)lDrug-druginteraction2021/9/13Hangzhou50Cohort studies in pharmacoepidemiologylStrengthslRelativeriskandAbsoluteriskestimateslMultipleoutcomes(safetysurveillanceformultipleout

38、comesandrisk-benefitassessment)lLimitationslMitigatedbytheavailabilityofsecondaryandhybriddatasources2021/9/13Hangzhou51Clinical trialslClinicalRegimentrialsvs.ClinicalpracticetrialslInclusionandExclusioncriterialTitration,switching,andstoppinglFollow-up/monitoringlNewDrugApplicationvs.Clinicalpract

39、icelPlacebo-controlledvs.head-to-headcomparisonlInternalvalidityvs.GeneralizabilitylComorbidityandpotentialdrug-druginteraction2021/9/13Hangzhou52Explanatory vs.PragmaticRolandandTorgerson.Understandingcontrolledtrials:Whatarepragmatictrials?BMJ1998;316:285“Trialsofhealthcareinterventionsareoftendes

40、cribedaseitherexplanatoryorpragmatic.Explanatorytrialsgenerallymeasureefficacythebenefitatreatmentproducesunderidealconditions,oftenusingcarefullydefinedsubjectsinaresearchclinic.Pragmatictrialsmeasureeffectivenessthebenefitthetreatmentproducesinroutineclinicalpractice.”2021/9/13Hangzhou53Some large

41、 trialslALLHATstudylhttp:/allhat.uth.tmc.edu/lPROVE-IT/TIMI-22studylhttp:/content.nejm.org/cgi/content/abstract/350/15/1495lCATIElhttp:/www.nimh.nih.gov/press/catie_release.cfmlhttp:/content.nejm.org/cgi/content/abstract/353/12/12092021/9/13Hangzhou54Clinical trials for safetylLargeandSimpleTriallLa

42、rgeandDirtytrial(Meinert)lAhybridapproachlRandomizationlIndividuallevellGrouplevellReal-lifemanagementandfollow-uplImplicationsforcomparativeeffectivenessresearch2021/9/13Hangzhou55Exampleofalargeandsimpletriallhttp:/clinicaltrials.gov/ct2/show/NCT00418171lLargeSimpleTrial(LST)OfCardiovascularSafety

43、OfZiprasidoneAndOlanzapine-(ZODIAC)lAnInternational,Multicenter,LargeSimpleTrial(LST)ToCompareTheCardiovascularSafetyOfZiprasidoneAndOlanzapinelTheprimaryobjectiveofthestudyistoestimatetherelativeincidenceamongusersofziprasidoneandolanzapineofnon-suicidemortalitywithin12months.l18,239enrolledpatient

44、s2021/9/13Hangzhou56Eligibility criteria for ZODIAClInclusionCriteria:lPatientsnewlytreatedforschizophreniaandthosereceivingcontinuingtreatmentwillbeeligibleifthetreatingpsychiatristisreadytoinitiateanewantipsychoticmedication,andwouldconsiderusingeitherziprasidoneorolanzapineasanappropriatetherapy.

45、lAge18andolderlMaleandfemalepatientswhomeetallcriterialistedbelowareeligibletobeenrolledinthisstudy:lDiagnosedwithschizophrenialWillingtoprovideinformationonatleastonealternatecontactpersonforstudystafftocontactregardingpatientswhereabouts,shouldthepatientbelost-to-follow-upoverthecourseofthestudylE

46、xclusionCriteria:lProgressivefataldiseaseofalifeexpectancywhichprohibitsthemfromparticipatinginaoneyearresearchstudylPreviouslyrandomizedtostudymedicationandenrolledinthisstudy2021/9/13Hangzhou57Clinical trial,statistical perspectivelOne(andonlyone)pre-specifiedprimaryendpointlTypeIandTypeIIerrors,s

47、amplesizeestimationlErronwhichside?lSubgroupanalysislHypothesisgeneration,notconfirmationlSuperioritytrialvs.Non-inferioritytriallImplicationsforthechoiceofcomparisongrouplMultiplecomparison(no-no!)2021/9/13Hangzhou58Subgroup analysislHortonR.Fromstarsignstotrialguidelines.Lancet2000;355:1033-34lISI

48、S-2Collaborationgroup.Randomisedtrialofintravenousstreptokinase,oralaspirin,both,orneitheramong17,187casesofsuspectedmyocardialinfarction.Lancet1988;ii:39-60l“TheanalysisshowedthataspirinwasbeneficialinallpatientsexceptthosewiththestarsignsofLibraandGemini.”2021/9/13Hangzhou59Differencebetweenpharma

49、coepidemiolgyandothertopicsinepidemiologylFocuslExposurevs.OutcomelConfounderslBiaslAgreementwithclinicaltrialslRegulatoryimplications2021/9/13Hangzhou60Whatiftheclinicaltrialsresultsdifferfromtheobservationalfindings?lWomensHealthInitiative,hormonereplacementtherapy,andcardiovascularoutcomeslhttp:/

50、jama.ama-assn.org/cgi/content/abstract/288/3/321lhttp:/jama.ama-assn.org/cgi/content/abstract/297/13/1465lhttp:/content.nejm.org/cgi/content/abstract/356/25/25912021/9/13Hangzhou61Clinical trials and observational studiesl“Rethinkingthehierarchyofevidence”VandenbrouckelObjectivesofstudylIntendedeffe