【文档说明】循证医学EvidenceBasedClinicalPracticeZSUFS课件.pptx，共(63)页，2.445 MB，由小橙橙上传

转载请保留链接：https://www.ichengzhen.cn/view-254967.html

以下为本文档部分文字说明：

Evidence-BasedMedicine循证医学杨学宁XueningYang,MD广东省人民医院广东省肺癌研究所广东省循证医学科研中心循证医学杂志社Evidence-basedPractice(EBP)Models•Tra

ditionalpracticeModel(×)•Model1:Usethehighestqualityinformationtoguideclinicaldecisions•Model2:Search,evalua

te,andmakeavailablespecialtyspecificbestavailableevidence•Model3:CreateoriginalresearchorsystematicreviewsEBMPractice-FivestepsPatientHi

storyandPE……•QuestionAskingAnswerableClinicalQuestions.•SearchSearchingforthebestcurrentevidence.•AppraiseCritically-appraisingtheevidenc

e.•Application:Integratingtheevidencewithourexpertiseandourpatient’suniquebiologyandvalues•AuditEvaluatingourperfor

manceStep1AskingAnswerableClinicalQuestions•Setting:64residentsat2NewHavenhospitals•Method:Interviewedafter401consultatio

ns•Questions–Asked280questions(2per3patients)–Pursuedananswerfor80questions(29%)–Notpursuedbecause•Lac

koftime•Forgotthequestion•Sourcesofanswers–Textbooks(31%),articles(21%),consultants(17%)Green,AmJMed2000Step1AskingAnswerableClinica

lQuestions•Mostofourquestionsareneveranswered•Whenanswered,theinformationislikelytobeneitherthebestnorup-to-dateStep1AskingAnswerableClini

calQuestions•EBMstartswithaclearclinicalquestion•Well-builtclinicalquestionsguidethewholeEBMprocess•Typeofquestion?–Causatio

n/etiology–Intervention–Diagnosis–Prognosis•TypeofStudy?–Whatwouldbethebeststudydesigninordertoanswerthequesti

on?–studydesignsshouldlimitbiasStep1AskingAnswerableClinicalQuestions•Questioncomponents:PICO–WhattypesofParticipants?–WhattypesofInterve

ntions?–WhattypesofComparison?–WhattypesofOutcomes?Step1AskingAnswerableClinicalQuestions•Whattypesofparticipants?–Diseaseorco

nditionofinterest–Potentialco-morbidity–Setting–Demographicfactors–TaskofclinicalproblemStep1AskingAnswerableClinicalQuest

ions•Whattypesofintervention?–Typesoftreatment–Typeofdiagnostictest–Typeofcausativeagent–TypeofprognosticfactorStep1AskingAnswerableClinica

lQuestions•Whattypesofoutcomes?–Forcausation,outcomeisdisease–Fordiagnostictest,outcomeindiseaseofinterest–Fortreatmentincludealloutcom

esimportanttopeoplemakingdecisionstodefinesuccessoftherapy–Forprognosis,outcomeisthechosenendpointofthediseaseStep1Asking

AnswerableClinicalQuestions•Broadornarrowquestions–Broad•Dogefinitibimprovesurvivalinlungadenocarcinoma?–Narrow•Dogefinitibi

mprovesurvivalinfemallungadenocarcinomarefractorytofirst-linechemo?Step1AskingAnswerableClinicalQuestions•Causation–People-Expos

ure-Comparison-Outcomes•Doesheavysmokerhavehigherincidenceoflungcancer?–Prognosis:People-Exposure-Comparison-Outcomes•Islu

ngadenocarcinomawithhighlevelofCEAaremorelikelytohaverecurrentdisease?Step2Searchingforthebestcurrentevidenc

e•Textbook–Annuallyrevised–Usehighlevelevidences–Withreference•Guidelines•CochraneLibrary•MedlineNationalGuidelineClearinghouse

(http://www.guideline.gov/)Step2Toomuch?•即使是小的问题也有相关指南Step2Step2CochraneLibrary•Advantage–highestqualityevidencewe’lleverhaveontheef

fectivenessofhealthcare•Disadvantage–notyetmanysystemicreviews–necessarilyomitsthenewesttreatmentStep2EBPModels•Traditi

onalpracticeModel(×)•Model1:Usethehighestqualityinformationtoguideclinicaldecisions•Model2:Search,evaluate,andmakeavailablespecialtyspecificLevel1inf

ormation•Model3:CreateoriginalresearchorsystematicreviewsStep2PubMed/MEDLINE•http://www.ncbi.nlm.nih.gov/PubmedStep2Medline-PubMedAdvant

age–exhaustiveness,flexibilityofsearching–journalcoverage,currency(on-lineversions)–widespreadavailabilityandsupport

Disadvantage–havetodoyourownqualityfiltering–puttingtogethergoodsearchesisdifficult–gapsincoverage(med

ical,geographicalandlinguistic)Step2Step2Step2Step2Step2Step2Step2Searchingforthebestcurrentevidence•Forprimaryliterature–No

tjustabstract!–TrytogetFull-textStep3Critically-appraisingtheevidence•Valid?–Designandmethodology:logical?–Sample–Experim

entorMeasure:carriedoutinasoundwayandConvincible？Follow-uprate?–Analysis–Conclusion–Ethical?•Importan

ce?–Clinicalsignificance？–Cost-effective？•Applicable?–Willtheresultshelplocally?Step3Critically-appraisingtheevidence•Quickscreeningby

title,abstractandrelatemessage–Authors•Well-known？Pastrecord？Creditable？–Institute•Competencetodothis

study？Well-known？–Specialty•Hasorhadpublishedincorejournals?Yourpractice？Valuable？•GoforFull-textDiscriminat

ion？Rational？EBM？Step3Critically-appraisingtheevidenceTypeofQuestionSuggestedbesttypeofStudyTherapyRCT>cohort>casecontrol>caseseriesDiagn

osisProspective,blindcomparisontoagoldstandardEtiology/HarmRCT>cohort>casecontrol>caseseriesPrognosisCohortstudy>casecont

rol>caseseriesPreventionRCT>cohortstudy>casecontrol>caseseriesClinicalExamProspective,blindcomparisontogoldstandardCostEconomic

analysisRightmethodsiscrucialTheEvidencePyramidHierarchyofevidence:arrangesstudydesignsbytheirsuscept

ibilitytobiasPrimaryResearchExperimentalobservationalSecondaryResearch•Forquantitativestudies•ForqualitativestudiesNoconsensusabout

therelativerigourofdifferentmethodsSystematicreviewStep3Critically-appraisingtheevidenceImprovedsurvivalandsamplesizedControlSurvival（%）5%10%15%101

,290394134151,894544208202,444676328252,926788374PhaseIIIRCTheadtoheadComparesurvivalAdequatesamplesizeStep3Critically-appraisingt

heevidence•Measure–Objective：Labtest，Radiology–Subjective：pain，symptoms•Endpoint–alternative：relieveofsymptom,s

hrinkoftumor–GoldStandard：prolongsurvivalStep3Critically-appraisingtheevidence•Conclusion–Basedonthedata?–New?•Updated?

•Comparewithotherstudy•And,Whatnext?Step3Critically-appraisingtheevidence•Attention!–Don’ttrustitjustbecauseitisfromabigboss,meta-analy

sis,RCT,orguideline.–Evidenceshouldbeinconsistencewithknownmedicalscience(physiology,pathology,pharmacology,etc）Usageofscien

tificlanguageisnotnecessarilymeanscientificbasedScientificmethodscreditNotalltheevidencesarethesameimporta

nceHollingworth&Jarvik,Radiology2007;244:31-38Flowchartofcost-effectivenessateachlevelofTAhierarchyTechnic

alPerformanceDiagnosticPerformanceDiagnosticImpactTherapeuticImpactImpactonHealthStep4Application•Integratingtheevidencewithourexpertiseandourpa

tient’suniquebiologyandvalues–Yourquestion–Yourpatient–Yourexpertise–Yourhospital–Howto……Step5Evaluatingou

rperformance•Follow-up•Evaluate–Thepatientsbenefit……?•Next……•……Oragainanother5steps•CollectdataandwriteanarticleEBPMode

ls•TraditionalpracticeModel(×)•Model1:Usethehighestqualityinformationtoguideclinicaldecisions•Model2:Search,evaluate,andmakeavailablespecialtyspec

ificbestavailableevidence•Model3:CreateoriginalresearchorsystematicreviewsSelectiveEBP•Maynotneedtocarryoutall5steps•1.AskingAnswe

rableClinicalQuestions.•X2.Searchingforthebestcurrentevidence.•X3.Critically-appraisingtheevidence.•4.Application:Integratingtheevid

encewithourexpertiseandourpatient’suniquebiologyandvalues•5.EvaluatingourperformanceSelectiveEBP•3differentmodesofE

BP–Searching&appraising•providesE-Bcare,butisexpensiveintimeandresources–Searchingonly•much,quicker,andifcarriedou

tamongE-Bresources,canprovideE-Bcare–Replicatingthepracticeofexperts•quickest,butmaynotdistinguishevidence-basedfromego-bas

edrecommendationsSelectiveEBP•EvenfullyEBM-trainedcliniciansworkinall3modes–Searching&appraising•fortheproblems

Iencounterdaily.–Searchingonly•forproblemsIencounteronceamonth.–Replicating•forproblemsIencounteronceadecade(andcrossingmyfingers!).Case1•Male,Z

SUProfessor，62yearsold•Smallcelllungcancer,limiteddisease,chemofor6cycleswithcompleteresponse–Medicaloncologist：prophylacticc

ranialirradiation（PCI）notimporvesurvival,andwithAE！–Radio-oncologist：PCIimprovesurvival,andtolerable•Question:Why?ShouldthepatientbegivePCI?Impr

ovesurvival?Howmuch?Sideeffect？When?How?TextbookPublish20061.ChanceofSCLCbrainmet：50%-80%for2years2.1995，aMETA-a

nalysis：SCLCwithCR,PCIimprove3yearsurvival15%-21%，butnotanalysisthecomplication3.Reducebrainmetfrom30%→5

%☆Conclusion：SCLCwithCR,PCIisrecommendafter2cyclesofchemoComplications?case1Guideline&SR•NCCN–LungcancerandPro

phylacticcranialirradiation1–2002TheCochraneLibrary,Issue2•Prophylacticcranialirradiationsignificantlyimprovessurvivalanddisease-freesurvivalfor

patientswithsmallcelllungcancerincompleteremission.case1例1FromSecondarySource•PCIshouldstartearlyafterchemo•comp

lications：–notRCT–2studyreportnodifferenceinmentalorneurologicaldisorder,butonlystudypartofthestudypatients，4

0%and60%（1995，1997）•N=300•N=314–InoneRCT（N=46）,reportamentaldisorder（1997）case1PUBMED•"CranialIrradiation"[MESH]ANDlung

neoplasms[mh]•("CranialIrradiation/adverseeffects"[MESH]OR"CranialIrradiation/mortality"[MESH])ANDlungneoplasms[mh]☆25papers，3pape

rsafter2006，onrelatepaper（level5）例1PUBMED•2009,JCOpeer-review•Complications•Mentalandneurological•Increase

byconcurrentchemoirradiation•Increaseindoselevel≥40Gy，nomorethan30Gy，recommend20Gy–SupportPCIinSCLClimiteddi

seasewithchemoCR.例1Conclusion&Decision•PCIreducebrainmet,improvesurvival•WhenRT<30GY,rarecomplications•Morebenefitthanharm•R

ecommendation：PCI，rightnowcase1EventsLikeThis……Sanlubabyformulamilkpowdercontaminatewithmelamine•Jul.16,2008–

甘肃省卫生厅接电话报告，称今年该院收治的婴儿患肾结石病例明显增多，近几个月已达十几例，经了解均食用了三鹿牌配方奶粉。•Sept.11,2008–新华社报道：“高度怀疑”三鹿公司生产的三鹿牌婴幼儿配方奶粉受到三聚氰胺污染。case2Sanlubabyformulam

ilkpowdercontaminatewithmelamine•Sept.17,2008国务院有关部门负责人和科学家–国家质检总局局长李长江：……共检验了109家企业的491批次产品，有22家企业的69个

批次的产品检出了三聚氰胺，检出不合格产品的企业约为20％……需要指出的是，这次被检出含有三聚氰胺产品的企业，不是所有批次的婴幼儿奶粉都有问题，比如，伊利、蒙牛、雅士利、施恩等企业被抽检的产品有几十个批次，但有问题的只有几个批次。case2Sanlubabyformulam

ilkpowdercontaminatewithmelamine–卫生部陈竺部长：三聚氰胺……主要对膀胱和肾脏有影响，引发动物膀胱炎、膀胱结石、肾脏炎症等，但没有发现对人类有致癌作用。婴幼儿食用……出现不明原因的哭闹、呕吐、发热、尿液混浊、血尿、少尿或无尿等症状，应立即就近到医疗机构筛查

就诊。食用含有三聚氰胺的其他品牌婴幼儿奶粉，只有出现上述症状时，才需要立即到医院筛查就诊。–从2008年9月12日至17日8时，各地报告临床诊断患儿一共有6244例。其中，4917例患儿症状轻微，生命体征稳定，正在进行院外随访治疗或已经治愈。现仍留院观察治疗患儿有1327人。所有临床诊断病

历中，有158人发生过急性肾功能衰竭，占总病例数的2.5%。经有效治疗，目前158例重症患儿当中，已经有94人病情平稳或已经治愈出院。此外，回顾性的调查发现有3例死亡病例，其中甘肃省有2名，浙江有1名，不

包括在刚才提到的6244名诊断病例之中。case2TheEvidencePyramidHierarchyofevidence:arrangesstudydesignsbytheirsusceptibilitytobiasPrimaryResearchExperimenta

lobservationalSecondaryResearch•Forquantitativestudies•ForqualitativestudiesNoconsensusabouttherelativerigourofdiffere

ntmethodsSystematicreviewSanlubabyformulamilkpowdercontaminatewithmelamine•Celloranimalisnoequaltohuman•Butthereareanycorelation?•Noevidencedoes

notmeannoharm!•Howshouldweact?Summary•EBMisaLifelonglearningmodel–lifelong,self-directed,problem-basedlearning•Know-howinfinding,appraising,andapplyi

ngevidenceFurtherReadingonEvidenceBasedMedicine