a review of the application of propensity score methods

a review of the application of propensity score methods

ID:7282643

大?。?02.83 KB

頁數:24頁

時間:2018-02-10

a review of the application of propensity score methods_第1頁
a review of the application of propensity score methods_第2頁
a review of the application of propensity score methods_第3頁
a review of the application of propensity score methods_第4頁
a review of the application of propensity score methods_第5頁
資源描述:

《a review of the application of propensity score methods》由會員上傳分享,免費在線閱讀,更多相關內容在工程資料-天天文庫。

1、NIHPublicAccessAuthorManuscriptJClinEpidemiol.Authormanuscript;availableinPMC2007May1.NIH-PAAuthorManuscriptPublishedinfinaleditedformas:NIH-PAAuthorManuscriptNIH-PAAuthorManuscriptJClinEpidemiol.2006May;59(5):437±447.Areviewoftheapplicationofpropensi

2、tyscoremethodsyieldedincreasinguse,advantagesinspecificsettings,butnotsubstantiallydifferentestimatescomparedwithconventionalmultivariablemethodsTilStürmer,MD,MPH1,2,ManishaJoshi,MS,MSW1,RobertJ.Glynn,PhD,ScD1,2,JerryAvorn,MD1,KennethJ.Rothman,DrPH1,3

3、,andSebastianSchneeweiss,MD,ScD11DivisionofPharmacoepidemiologyandPharmacoeconomics,BrighamandWomen’sHospital,HarvardMedicalSchool,Boston,MA,USA2DivisionofPreventiveMedicine,BrighamandWomen’sHospital,HarvardMedicalSchool,Boston,MA,USA3DepartmentofEpid

4、emiology,BostonUniversitySchoolofPublicHealthAbstractObjective—Propensityscoreanalysesattempttocontrolforconfoundinginnon-experimentalstudiesbyadjustingforthelikelihoodthatagivenpatientisexposed.Suchanalyseshavebeenproposedtoaddressconfoundingbyindica

5、tion,butthereislittleempiricalevidencethattheyachievebettercontrolthanconventionalmultivariateoutcomemodeling.Studydesignandmethods—UsingPubMedandScienceCitationIndex,weassessedtheuseofpropensityscoresovertimeandcriticallyevaluatedstudiespublishedthro

6、ugh2003.Results—Useofpropensityscoresincreasedfromatotalof8papersbefore1998to71in2003.Mostofthe177publishedstudiesabstractedassessedmedications(N=60)orsurgicalinterventions(N=51),mainlyincardiologyandcardiacsurgery(N=90).WhetherPSmethodsorconventional

7、outcomemodelswereusedtocontrolforconfoundinghadlittleeffectonresultsinthosestudiesinwhichsuchcomparisonwaspossible.Only9outof69studies(13%)hadaneffectestimatethatdifferedbymorethan20%fromthatobtainedwithaconventionaloutcomemodelinallPSanalysespresente

8、d.Conclusions—Publicationofresultsbasedonpropensityscoremethodshasincreaseddramatically,butthereislittleevidencethatthesemethodsyieldsubstantiallydifferentestimatescomparedwithconventionalmultivariablemethods.Keywordspropensityscore;epidemiolo

當前文檔最多預覽五頁,下載文檔查看全文

此文檔下載收益歸作者所有

當前文檔最多預覽五頁,下載文檔查看全文
溫馨提示:
1. 部分包含數學公式或PPT動畫的文件,查看預覽時可能會顯示錯亂或異常,文件下載后無此問題,請放心下載。
2. 本文檔由用戶上傳,版權歸屬用戶,天天文庫負責整理代發(fā)布。如果您對本文檔版權有爭議請及時聯(lián)系客服。
3. 下載前請仔細閱讀文檔內容,確認文檔內容符合您的需求后進行下載,若出現內容與標題不符可向本站投訴處理。
4. 下載文檔時可能由于網絡波動等原因無法下載或下載錯誤,付費完成后未能成功下載的用戶請聯(lián)系客服處理。