如何解讀診斷性臨床試驗(文件)
【正文】 tage requirements accuracy high specificity fast ,convenient high sensitivity costs high low treatments treat when positive further tests when positive 復(fù)旦大學(xué)公共衛(wèi)生學(xué)院 Elements in design ?gold standard ? pathologic diagnosis(biopsy,autopsy) ? surgical discovery ? coronary arteriongraphy ? clinical follow up ?objects： ? patients： diagnosted by gold standard, representative, including the whole spectrum of the disease: all stages,types,severity,plications ? controls： including healthy people and people with other diseases 復(fù)旦大學(xué)公共衛(wèi)生學(xué)院 Elements in design ?Diagnostic indicator and judge criteria ? subjective indicator： chief plaint ? semiobjective indicator:judges from examiner ? objective indicator： indicators measured by instruments ?Blinding ?Sample size estimation 復(fù)旦大學(xué)公共衛(wèi)生學(xué)院 Validity Sensitivity： Se=a/(a+c) the proportion of people with the disease who have a positive test for the disease Specificity： Sp=d/(b+d) the proportion of people without the disease who have a negative test。 when the prevalence of population is %, then PPV=*(*+*)=% hence: a positive test result means the probability of real HIV(+) is %. PPV (1 ) (1 )S e PS e P S p P?? ? ? ? ?復(fù)旦大學(xué)公共衛(wèi)生學(xué)院 Result Reliability index ? kappa statistic: ? If Kappa = 1, then there is perfect agreement. ? If Kappa = 0, then there is no agreement. The higher the value of Kappa, the stronger the agreement For diagnostic tests: ? k ≥0. 75 good； ? 0. 4 k 0. 75 general； ? k ≤0. 4 bad 復(fù)旦大學(xué)公共衛(wèi)生學(xué)院 Tradeoffs between sensitivity and specificity Put weights to Se and Sp ?Same weights， then maximum Se+Sp use the ROC curve :the top left corner ? Different weights， then maximum α Se+Sp ?Considerations： ? further examinations ? consequences of misdiagnosis or missed diagnosis ? next examination ? prevalence ? necessity of treatment 復(fù)旦大學(xué)公共衛(wèi)生學(xué)院 Comparisons and statistic tests ?Area under the ROC curve Area should ? test H0:the area= , ? if H0 ,then the diagnostic test have no clinical value ?Different diagnostic tests the larger the area,the better the test 復(fù)旦大學(xué)公共衛(wèi)生學(xué)院 Considerations ?Costbenefit/effect analysis, CBA/ CEA ?Increasing th
研究報告相關(guān)推薦
鄂ICP備17016276號-1