Evidence-BasedMedicine

循證醫(yī)學(xué)楊學(xué)寧XueningYang,MD廣東省人民醫(yī)院廣東省肺癌研究所廣東省循證醫(yī)學(xué)科研中心循證醫(yī)學(xué)雜志社Evidence-basedPractice(EBP)ModelsTraditionalpracticeModel(×)Model1:UsethehighestqualityinformationtoguideclinicaldecisionsModel2:Search,evaluate,andmakeavailablespecialtyspecificbestavailableevidenceModel3:CreateoriginalresearchorsystematicreviewsEBMPractice-Fivesteps

PatientHistoryandPE……QuestionAskingAnswerableClinicalQuestions.SearchSearchingforthebestcurrentevidence.AppraiseCritically-appraisingtheevidence.Application:Integratingtheevidencewithourexpertiseandourpatient’suniquebiologyandvaluesAuditEvaluatingourperformanceStep1AskingAnswerableClinicalQuestionsSetting:64residents

at2NewHavenhospitalsMethod:Interviewedafter401consultationsQuestionsAsked280questions(2per3patients)Pursuedananswerfor80questions(29%)NotpursuedbecauseLackoftimeForgotthequestionSourcesofanswersTextbooks(31%),articles(21%),consultants(17%)Green,AmJMed2000Step1AskingAnswerableClinicalQuestionsMostofourquestionsareneveransweredWhenanswered,theinformationislikelytobeneitherthebestnorup-to-dateStep1AskingAnswerableClinicalQuestionsEBMstartswithaclearclinicalquestionWell-builtclinicalquestionsguidethewholeEBMprocessTypeofquestion?Causation/etiologyInterventionDiagnosisPrognosisTypeofStudy?Whatwouldbethebeststudydesigninordertoanswerthequestion?studydesignsshouldlimitbiasStep1AskingAnswerableClinicalQuestionsQuestioncomponents:PICOWhattypesofParticipants?WhattypesofInterventions?WhattypesofComparison?WhattypesofOutcomes?Step1AskingAnswerableClinicalQuestionsWhattypesofparticipants?DiseaseorconditionofinterestPotentialco-morbiditySettingDemographicfactorsTaskofclinicalproblemStep1AskingAnswerableClinicalQuestionsWhattypesofintervention?TypesoftreatmentTypeofdiagnostictestTypeofcausativeagentTypeofprognosticfactorStep1AskingAnswerableClinicalQuestionsWhattypesofoutcomes?Forcausation,outcomeisdiseaseFordiagnostictest,outcomeindiseaseofinterestFortreatmentincludealloutcomesimportanttopeoplemakingdecisionstodefinesuccessoftherapyForprognosis,outcomeisthechosenendpointofthediseaseStep1AskingAnswerableClinicalQuestionsBroadornarrowquestionsBroadDogefinitibimprovesurvivalinlungadenocarcinoma?NarrowDogefinitibimprovesurvivalinfemallungadenocarcinomarefractorytofirst-linechemo?Step1AskingAnswerableClinicalQuestionsCausationPeople-Exposure-Comparison-OutcomesDoesheavysmokerhavehigherincidenceoflungcancer?Prognosis:People-Exposure-Comparison-OutcomesIslungadenocarcinomawithhighlevelofCEAaremorelikelytohaverecurrentdisease?Step2SearchingforthebestcurrentevidenceTextbookAnnuallyrevisedUsehighlevelevidencesWithreferenceGuidelinesCochraneLibraryMedlineNationalGuidelineClearinghouse(:///)Step2Toomuch?即使是小的問題也有相關(guān)指南Step2Step2CochraneLibraryAdvantagehighestqualityevidencewe’lleverhaveontheeffectivenessofhealthcareDisadvantagenotyetmanysystemicreviewsnecessarilyomitsthenewesttreatmentStep2EBPModelsTraditionalpracticeModel(×)Model1:UsethehighestqualityinformationtoguideclinicaldecisionsModel2:Search,evaluate,andmakeavailablespecialtyspecificLevel1informationModel3:CreateoriginalresearchorsystematicreviewsStep2PubMed/MEDLINE://Step2Medline-PubMedAdvantageexhaustiveness,flexibilityofsearchingjournalcoverage,currency(on-lineversions)widespreadavailabilityandsupportDisadvantagehavetodoyourownqualityfilteringputtingtogethergoodsearchesisdifficultgapsincoverage(medical,geographicalandlinguistic)Step2Step2Step2Step2Step2Step2Step2SearchingforthebestcurrentevidenceForprimaryliteratureNotjustabstract!TrytogetFull-textStep3Critically-appraisingtheevidenceValid?Designandmethodology:logical?SampleExperimentorMeasure:carriedoutinasoundwayandConvincible？Follow-uprate?Analysis

ConclusionEthical?Importance?Clinicalsignificance？Cost-effective？Applicable?Willtheresultshelplocally?Step3Critically-appraisingtheevidenceQuickscreeningbytitle,abstractandrelatemessageAuthorsWell-known？Pastrecord？Creditable？InstituteCompetencetodothisstudy？Well-known？SpecialtyHasorhadpublishedincorejournals?Yourpractice？Valuable？GoforFull-textDiscrimination？Rational？EBM？Step3Critically-appraisingtheevidence

TypeofQuestionSuggestedbesttypeofStudyTherapyRCT>cohort>casecontrol>caseseriesDiagnosisProspective,blindcomparisontoagoldstandardEtiology/HarmRCT>cohort>casecontrol>caseseriesPrognosisCohortstudy>casecontrol>caseseriesPreventionRCT>cohortstudy>casecontrol>caseseriesClinicalExamProspective,blindcomparisontogoldstandardCostEconomicanalysisRightmethodsiscrucialTheEvidencePyramidHierarchyofevidence:arrangesstudydesignsbytheirsusceptibilitytobiasPrimaryResearch

Experimental

observationalSecondaryResearchForquantitativestudiesForqualitativestudiesNoconsensusabouttherelativerigourofdifferentmethodsSystematicreviewStep3Critically-appraisingtheevidencePhaseIIIRCTheadtoheadComparesurvivalAdequatesamplesizeStep3Critically-appraisingtheevidenceMeasureObjective：Labtest，RadiologySubjective：pain，symptomsEndpointalternative：relieveofsymptom,shrinkoftumorGoldStandard：prolongsurvivalStep3Critically-appraisingtheevidenceConclusionBasedonthedata?New?Updated?ComparewithotherstudyAnd,Whatnext?Step3Critically-appraisingtheevidenceAttention!Don’ttrustitjustbecauseitisfromabigboss,meta-analysis,RCT,orguideline.Evidenceshouldbeinconsistencewithknownmedicalscience(physiology,pathology,pharmacology,etc〕UsageofscientificlanguageisnotnecessarilymeanscientificbasedScientificmethodscreditNotalltheevidencesarethesameimportanceHollingworth&Jarvik,Radiology2007;244:31-38Flowchartofcost-effectivenessateachlevelofTAhierarchyTechnicalPerformanceDiagnosticPerformanceDiagnosticImpactTherapeuticImpactImpactonHealthStep4ApplicationIntegratingtheevidencewithourexpertiseandourpatient’suniquebiologyandvaluesYourquestionYourpatientYourexpertiseYourhospitalHowto……Step5EvaluatingourperformanceFollow-upEvaluateThepatientsbenefit……?Next…………Oragainanother5stepsCollectdataandwriteanarticleEBPModelsTraditionalpracticeModel(×)Model1:UsethehighestqualityinformationtoguideclinicaldecisionsModel2:Search,evaluate,andmakeavailablespecialtyspecificbestavailableevidenceModel3:CreateoriginalresearchorsystematicreviewsSelectiveEBPMaynotneedtocarryoutall5steps1.AskingAnswerableClinicalQuestions.X2.Searchingforthebestcurrentevidence.X3.Critically-appraisingtheevidence.4.Application:Integratingtheevidencewithourexpertiseandourpatient’suniquebiologyandvalues5.EvaluatingourperformanceSelectiveEBP3differentmodesofEBPSearching&appraisingprovidesE-Bcare,butisexpensiveintimeandresourcesSearchingonlymuch,quicker,andifcarriedoutamongE-Bresources,canprovideE-BcareReplicatingthepracticeofexpertsquickest,butmaynotdistinguishevidence-basedfromego-basedrecommendationsSelectiveEBPEvenfullyEBM-trainedcliniciansworkinall3modesSearching&appraisingfortheproblemsIencounterdaily.SearchingonlyforproblemsIencounteronceamonth.ReplicatingforproblemsIencounteronceadecade(andcrossingmyfingers!).Case1Male,ZSUProfessor，62yearsoldSmallcelllungcancer,limiteddisease,chemofor6cycleswithcompleteresponseMedicaloncologist：prophylacticcranialirradiation〔PCI〕notimporvesurvival,andwithAE！Radio-oncologist：PCIimprovesurvival,andtolerableQuestion:Why?ShouldthepatientbegivePCI?Improvesurvival?Howmuch?Sideeffect？When?How?TextbookPublish20061.ChanceofSCLCbrainmet：50%-80%for2years2.1995，aMETA-analysis：SCLCwithCR,PCIimprove3yearsurvival15%-21%，butnotanalysisthecomplication3.Reducebrainmetfrom30%→5%☆Conclusion：SCLCwithCR,PCIisrecommendafter2cyclesofchemoComplications?case1Guideline&SRNCCNLungcancerandProphylacticcranialirradiation12002TheCochraneLibrary,Issue2Prophylacticcranialirradiationsignificantlyimprovessurvivalanddisease-freesurvivalforpatientswithsmallcelllungcancerincompleteremission.

case1例1FromSecondarySourcePCIshouldstartearlyafterchemocomplications：notRCT2studyreportnodifferenceinmentalorneurologicaldisorder,butonlystudypartofthestudypatients，40%and60%〔1995，1997〕N=300N=314InoneRCT〔N=46〕,reportamentaldisorder〔1997〕case1PUBMED"CranialIrradiation"[MESH]ANDlungneoplasms[mh]("CranialIrradiation/adverseeffects"[MESH]OR"CranialIrradiation/mortality"[MESH])ANDlungneoplasms[mh]☆25papers，3papersafter2006，onrelatepaper〔level5〕例1PUBMED2021,JCOpeer-reviewComplicationsMentalandneurologicalIncreasebyconcurrentchemoirradiationIncreaseindoselevel≥40Gy，nomorethan30Gy，recommend20GySupportPCIinSCLClimiteddiseasewithchemoCR.例1Conclusion&DecisionPCIreducebrainmet,improvesurvivalWhenRT<30GY,rarecomplicationsMorebenefitthanharmRecommendation：PCI，rightnowcase1EventsLikeThis……

SanlubabyformulamilkpowdercontaminatewithmelamineJul.16,2021甘肅省衛(wèi)生廳接報告，稱今年該院收治的嬰兒患腎結(jié)石病例明顯增多，近幾個月已達十幾例，經(jīng)了解均食用了三鹿牌配方奶粉。Sept.11,2021新華社報道：“高度疑心〞三鹿公司生產(chǎn)的三鹿牌嬰幼兒配方奶粉受到三聚氰胺污染。case2SanlubabyformulamilkpowdercontaminatewithmelamineSept.17,2021國務(wù)院有關(guān)部門負責(zé)人和科學(xué)家國家質(zhì)檢總局局長李長江：……共檢驗了109家企業(yè)的491批次產(chǎn)品，有22家企業(yè)的69個批次的產(chǎn)品檢出了三聚氰胺，檢出不合格產(chǎn)品的企業(yè)約為20％……需要指出的是，這次被檢出含有三聚氰胺產(chǎn)品的企業(yè)，不是所有批次的嬰幼兒奶粉都有問題，比方，伊利、蒙牛