08-4影响临床科研质量的因素

–––(validity)(validity)–(internalvalidity)–externalvalidity)–()(reliability)(reliability)–•••–––––()––––——()(systematicerror)(bias)–•–••••–•–––••––“”––Meittinen(1976)–(selectionbias)–(informationbias)–(confoundingbias)–•1(admissionratebias)–BX•BB1000•20X•BX•B50%•B20%•X40%X2=15.215,p0.01OR=1.86––––2.(misclassificationbias)––––3.(detectionsignbias)––110–1979HorwitFeintein••••––Horwitz:–7958–:4.(non-respondentbias)–––Taylor1966––73.658.2–4324––––•6520–––••–Alderson90Sackett80805.(volunteerbias)–:–––––,,––6.(withdrawbiaslossoffollow-up)––7.(prevalence-incidencebias)–(Neymanbias)––––––Friedman–RR2.4–OR1.16–42.325.1––8.(susceptibilitybias)––:1971Stern:–––:–9.(membershipbias)–––••10(exclusivebias)––––11(noncontemporarybias)–––––13(diagnosticenterbias)–1–––––2––2––––••3–••–•4––••10–••–•••••(measurementbias)––•••1(diagnosticsuspicionbias––––––––––(FraminghamHeartStudy)––––38–2(exposuresuspicionbias)––•••••1962NishiyamaKaventos–280–47503(recallbias)––•••••1956StewartX–XX–X–X–/–X–4(missingclinicaldatabias)––5(familyinformationbias)–Schull–––6sequencebias):–7(unreceptivemeasurebias)–8(non-sensitivebias)––1234–(extraneousfactor)–(confoundingfactor)–••••–•603660•602440–––––605287–60813–1––•257A17869•514A24347•(cOR)1.50AB–(aOR)1–——2––•60020033•5400180033•cOR1––•5050(aOR)1–1(restriction)2(matching)3(randomization)4(stratification)5.(standardization)6(multivariateanalysis)(chance)–(randomerror)(chance)––••••–––––––––