観察研究の必須事項

  • Published on
    11-Nov-2014

  • View
    579

  • Download
    2

Embed Size (px)

DESCRIPTION

http://t.co/r7okiho781 78 http://jpa2014.com/ SS-067 2014/09/11 15:30-17:30 () (): (): () : (): (): ():

Transcript

  • 1. 78SS-0672014911RY302

2. 84(Annals of Internal Medicine, Lancet)100STROBEAnn Intern Med. 2007;147:W-163W-1942& 3. (observational study)(cohort study)(case control study)(cross sectional study)() 4. 41++ RCTRCT1+ , RCT, RCT1- , RCT, RCT2++ 2+ 2- 3 4 AHRQ: http://www.guideline.gov/content.aspx?id=39324 5. 5:= +(RE): random error(SE): systematic error=RESERESERESERESE 6. STROBE1(I) 2(M) 81(R) 23(D) 41184226 7. 7 8. 8 9. 9 10. 10 11. (null=true) (null=false)(=0)(1-)(0)()(1-).80(Cohen, 1992) 12. 13. = R2 (.01-.05)=1- (.80-.95)= (.05)0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5180160140120100806040200Total sample sizePower (1- err prob)= 0.95= 0.9= 0.85= 0.8F tests - Linear mult iple regression: Fixed model. R deviat ion f rom zeroNumber of predictors = 3. err prob = 0.05Effect size f 14. 14 15. 151) 2) 3) 4) ( )( )( )( ) 16. (99%, 95%, 90%) ( .30, .50, .90)()Kelly & Rauch. Psychological Methods 2006,Vol. 11, No. 4, 363385 17. 17 18. 18 19. 19> validate> validate2 20. 20() Vet, H. C. W., Terwee, C. B.,Mokkink, L. B., & Knol, D.J. (2011) Measurement inmedicine. A practical guide.Cambridge: CambridgeUniversity Press.http://www.slideshare.net/yoshitaket/cosmin-takebayashi-21816803 21. 21 22. 22 23. 23Self-selection biasHealthy worker effectAdmission biasIncidence-prevalence biasWithdrawals biasNon-respondent bias 24. 24Diagnostic biasRecall biasRumination biasInterviewer biasMeasurement biasMisclassification bias 25. ()- - - - 25 26. = 26 27. 27 28. or 28= 29. 29: 30. 30XX 31. Take Home Message1. 2. 3. 4. 31 32. Contact to : ytake2 ism.ac.jp32https://sites.https://sites.google.com/site/yoshitake2bayashi/ google.com/site/studygroup13csrm/activity