诊断试验证实偏倚校正方法的比较研究

目的探讨两阶段两分类诊断试验存在证实偏倚时灵敏度和特异度的校正方法并进行比较。方法运用实例和模拟数据说明传统方法灵敏度和特异度的有偏估计,并运用最大似然估计和贝叶斯估计对灵敏度和特异度进行校正。结果传统方法计算的灵敏度和特异度存在偏倚,最大似然法和贝叶斯方法虽然均可校正估计偏倚,但是后者计算的可信区间比前者窄。结论在对诊断试验证实偏倚进行校正时,若有良好的先验信息,贝叶斯方法所获的结果更精确。

Comparison the Correcting for Verification Bias in Studies of a Diagnostic Test

ObjectiveCorrecting sensitivity and specificity in studies of a two-phase single binary diagnostic test when verification bias is present.MethodsComparing the estimate of sensitivity and specificity by traditional,frequency and Bayesian method.The estimators are illustrated using data from reviewing issues and simulating the data.ResultsBias is present when using traditional method to estimate sensitivity and specificity.Frequency and Bayesian method can correct verification bias.However,the confidence interval (CI) of the Bayesian method is smaller than the frequency method.ConclusionIf there is a good prior information,the results from Bayesian method will much more accuracy in a diagnostic test when verification bias is present.