意思是使恢复健康讲解.docx

意思是使恢复健康讲解.docx

《意思是使恢复健康讲解.docx》由会员分享，可在线阅读，更多相关《意思是使恢复健康讲解.docx（14页珍藏版）》请在冰豆网上搜索。

意思是使恢复健康讲解

Type2diabetesmellitusaftergestationaldiabetes:

asystematicreviewandmeta-analysis

LeanneBellamy,Juan-PabloCasas,AroonDHingorani,DavidWilliams

Lancet2009;373:

1773–79

Introduction

Gestationaldiabetesmellitusisdefinedasglucoseintolerancethatisfirstdetectedduringpregnancy.IntheUSA,thisconditionaffects135000pregnancies（3–5%ofallpregnancies）peryear.Shortlyafterdelivery,glucosehomoeostasisisrestoredtonon-pregnancylevels,butaffectedwomenremainathighriskofdevelopingtype2diabetesmellitusinthefuture.

Foranypopulationandethnicgroup,theriskofgestationaldiabetesindicatestheunderlyingfrequencyoftype2diabetes.Theincidencesofgestationaldiabetesandtype2diabetesarerisingthroughouttheworld,withhugehealth-careandeconomiccosts.Diabetespredisposesindividualstocardiovascular,renal,andretinaldiseases,costingUS$91·8billionperyearintheUSA.（A）

Womenwhohavehadgestationaldiabetesareadvisedtohavetheirglucosetoleranceassessed6weeksafterdelivery.However,lowratesofattendanceatthe6-weekfollow-upsuggestthathealth-careprofessionals,womenwithgestationaldiabetes,orboth,donotrealizetheimportanceofthisdisorderasanearlywarningsignofthesusceptibilitytodeveloptype2diabetesinthefuture;thereforeanopportunitytopromotehealthandpreventdiseaseismissed.Moreover,noconsensusexistsonhowandwhethermothersshouldcontinuetobemonitoredafterthisperiod.（B）

Theassociationbetweengestationaldiabetesandtype2diabetesmellitushasimplicationsfortheelucidationofthecausesofthesedisorders,andforthepredictionandpossiblepreventionordelayofthedevelopmentoftype2diabetesinwomen.（C）Wethereforedidacomprehensivesystematicreviewandmetaanalysistoquantifytheoverallriskofwomenwithgestationaldiabetesmellitusdevelopingtype2diabetesmellitus,andtoassesstheeffectoffactorsthatmightmodifythisrisk.（D）

Stockphrases:

1.restore

意思是“使恢复健康，使复原”为不及物动词，如后面接宾语+to

Hewascompletelyrestored.

我们熟悉的恢复健康一般用recover，但是recover主要是说从疾病状态恢复健康，而restore有从abnormal但是并非sick的状态恢复正常，用这个更加确切。

2.predispose

v.使易罹患,使偏向于

一般会想到makesusceptible，但是predispose更偏重于易于患病。

Theinhabitantsarepredisposedtorheumatismbythedampclimate.

3.assess

是评估评价较常用的词

4.promote

促进健康用promotehealth，中promote为动词，如果做健康促进解应用healthpromotion

5.noconsensusexistson

在某些方面意见还不统一的地道用法

6.theelucidationof意味解释阐明，具有挖掘原因说明原理之意。

比explain更具有科学味道。

7.modify用在本文中意为影响之意，较affect更谨慎。

Equivalenceoftwotreatmentsandsamplesizedeterminationunderexponentialsurvivalmodelwithcensoring

Jun-moNama,JinheumKimb,SeungyeounLeec,∗

ComputationalStatistics&DataAnalysis49（2005）217–226

Recently,therehasbeenagreatinterestinestablishingequivalenceoftwotreatments

inclinicaltrials.Forexample,astandardchemotherapyinpediatriconcologyishighly

effectivebutcausesseveretoxicside-effects,andresearchersareinterestedinalesstoxic

newtreatmentwhichmaybeessentiallyaseffectiveasthestandardoneinsurvival（Patte

etal.,1991）.Theintentionofanequivalencetrialistodemonstratethattwotreatments

donotdifferbymorethanaprescribedsmallamountwhichismateriallyinsignificant.

Theconventionaltestprocedurefordetectingadifferenceinacomparativetrialcannotbe

appliedforthissituation.

Statisticalmethodsforestablishingone-sidedequivalenceornon-inferiorityofanew

treatmenttothestandardoneonbinaryresponseshavebeeninvestigatedbymanyauthors,

e.g.,DunnettandGent（1977）,RoebruckandKühn（1995）andNam（1997）.Forequivalence

oftwosurvivaldistributionswithcensoredobservations,Wellek（1993）andCom-Nougue

etal.（1993）haveproposedtestingproceduresbasedontheproportionalhazardsmodel.

Wellek（1993）derivedtheuniformlymostpowerfultestintermsofthemaximumpartial

likelihoodestimatorbutthesamplesizeequationisnotgivenexplicitlywhileCom-Nougue

etal.（1993）providedtheconfidenceintervalsfortheactualhazardratiobasedonthelog

rankteststatistic.Whendatafollowanexponentialmodelwithnocensoring,Bristoland

Desu（1990）havesuggestedaparametricmethodoftestingforequivalence.（A）However,aparametricmethodbasedoncensoreddatahasnotbeenthoroughlystudied.（B）

Inthispaper,weinvestigatestatisticalmethodsinvolvingtheequivalenceoftwotreat-

mentsbasedonexponentiallydistributedsurvivaldatawithcensoring.（C）InSection2,we

derive（比propose好，显得是根据推理得来的）twodifferenttestsforequivalence:

thescoretestandFtestprocedures.Inaddition,

theasymptoticpowersandapproximatesamplesizeformulaareprovided.InSection3,the

scoreandFtestsarecomparedwiththelogranktest（Com-Nougueetal.,1993）bysimula-

tionsinlevelandpower,andapproximatenumbersofeventsrequiredforaspecificpower

usingthesemethodsareexamined.Also,weinvestigatetherobustnessofthethreetests

whentheunderlyingexponentialmodelisviolated.Sections4and5containanexample（好句）basedonnon-Hodgkin’smalignanttypeBlymphomadataanddiscussion.（D）

Somefundamentalissueswithnon-inferioritytestinginactivecontrolledtrials‡

H.M.JamesHung1;∗;†,Sue-JaneWang2,YiTsong3,JohnLawrence1andRobertT.O’Neil4

STATISTICSINMEDICINE2003;22:

213–225

Non-inferioritytestingfortherapeuticeffectivenessinactivecontrolledclinicaltrialswith-

outaplaceboarmisoftencontroversialdespitethefactthatthereisconsiderablepractical

experienceinusingthisapproachtoestablishtheeffectivenessofanexperimentaltherapy.

Thecontroversysurrounds（比contain好）manyissuesincludingthechoiceofactivecontrol,theselectionofhistoricaltrialstoestimatetheeffectoftheactivecontrol,statisticalmethodsforestimationofthecontroleffect,methodsfordeterminationofthenon-inferioritymargin,theanalysispopulationetc.[1–16].（A）

Manyotherfundamentalissueshavenotbeenaddressed（比soloved好）inpracticalapplicationsofnon-inferioritytesting.Oftenthemainobjectiveofnon-inferioritytestingisnotclear.Determinationofanon-inferioritymarginisalwaysunderheateddebate.Tomakestatisticalinferencefornon-inferioritytesting,therelevantalphaerror（ortypeIerror）probabilityofmakingafalseconclusionisnotclearlydefined.（B）Inthispaperwewillexploretheseissuesingreatdetail.（C）InSection2wewilldiscussthattheobjectiveofthenon-inferiorityanalysisisthekeytotheformulationoftheintendedstatisticalhypothesisandtothelevelofcomplexityindefiningthenon-inferioritymargin.Section3isdevotedtotheissuespertainingtothedefinitionofthenon-inferioritymargin.Weshallexposethepotentialproblemsindefiningthenon-inferioritymarginbasedontheestimateoftheactivecontrolef-

fect.InSection4,twodifferenttypesofstatisticalmethodsforthenon-inferiorityanalysiswillbediscussed.Alphaerrorprobabilitywillbeinvestigated.Section5givesanexampletoexplorethealphaerrorsofthetwotypesofstatisticalmethods.Concludingdiscussion

follows.（D）

Equivalenceoftwotreatmentsandsamplesizedeterminationunderexponentialsurvivalmodelwithcensoring

Jun-moNama,JinheumKimb,SeungyeounLeec,∗

ComputationalStatistics&DataAnalysis49（2005）217–226

Recently,therehasbeenagreatinterestinestablishingequivalenceoftwotreatments

inclinicaltrials.Forexample,astandardchemotherapyinpediatriconcologyishighly

effectivebutcausesseveretoxicside-effects,andresearchersareinterestedinalesstoxic

newtreatmentwhichmaybeessentiallyaseffectiveasthestandardoneinsurvival（Patte

etal.,1991）.Theintentionofanequivalencetrialistodemonstratethattwotreatments

donotdifferbymorethanaprescribedsmallamountwhichismateriallyinsignificant.

Theconventionaltestprocedurefordetectingadifferenceinacomparativetrialcannotbe

appliedforthissituation.

Statisticalmethodsforestablishingone-sidedequivalenceornon-inferiorityofanew

treatmenttothestandardoneonbinaryresponseshavebeeninvestigatedbymanyauthors,

e.g.,DunnettandGent（1977）,RoebruckandKühn（1995）andNam（1997）.Forequivalence

oftwosurvivaldistributionswithcensoredobservations,Wellek（1993）andCom-Nougue

etal.（1993）haveproposedtestingproceduresbasedontheproportionalhazardsmodel.

Wellek（1993）derivedtheuniformlymostpowerfultestintermsofthemaximumpartial

likelihoodestimatorbutthesamplesizeequationisnotgivenexplicitlywhileCom-Nougue

etal.（1993）providedtheconfidenceintervalsfortheactualhazardratiobasedonthelog

rankteststatistic.Whendatafollowanexponentialmodelwithnocensoring,Bristoland

Desu（1990）havesuggestedaparametricmethodoftestingforequivalence.（A）However,aparametricmethodbasedoncensoreddatahasnotbeenthoroughlystudied.（B）

Inthispaper,weinvestigatestatisticalmethodsinvolvingtheequivalenceoftwotreat-

mentsbasedonexponentiallydistributedsurvivaldatawithcensoring.（C）InSection2,we

derive（比propose好，显得是根据推理得来的）twodifferenttestsforequivalence:

thescoretestandFtestprocedures.Inaddition,

theasymptoticpowersandapproximatesamplesizeformulaareprovided.InSection3,the

scoreandFtestsarecomparedwiththelogranktest（Com-Nougueetal.,1993）bysimula-

tionsinlevelandpower,andapproximatenumbersofeventsrequiredforaspecificpower

usingthesemethodsareexamined.Also,weinvestigatetherobustnessofthethreetests

whentheunderlyingexponentialmodelisviolated.Sections4and5containanexample（好句）basedonnon-Hodgkin’smalignanttype