无金标准条件下含协变量的ROC曲线估计方法  

Estimating Receiver Operating Characteristic Curves with Covariates in the Absence of Gold Standard Test

作　　者：刘妍[1] 王红宇[2] 田娇妮[1] 刘桂芬[1] 

机构地区：[1]山西医科大学卫生统计教研室,030001 [2]山西医科大学第二医院

出　　处：《中国卫生统计》2012年第3期321-324,共4页Chinese Journal of Health Statistics

基　　金：国家自然科学基金项目(编号81172774);山西省自然科学基金项目(编号2009011005-2)

摘　　要：目的阐明无金标准条件下,考虑协变量后估计ROC曲线的两部贝叶斯模型。方法介绍两部贝叶斯模型,结合实例,筛选无金标准条件下ROC曲线的影响因素,考虑协变量影响后,估计ROC曲线。结果两部贝叶斯模型不仅可探讨协变量对疾病状态的影响,而且可探讨协变量对诊断试验结果的影响,同时可计算不同协变量取值条件下ROC曲线下面积。结论两部贝叶斯模型可有效地解决无金标准条件下,考虑协变量影响的ROC曲线估计问题。Objective To introduce the method for estimation of receiver operating characteristic curves with covariates when there is no gold standard test. Methods To estimate the ROC curve after considered the effects of covariates, using two-part Bayesian model screening the im- pact factors for the ROC curve without gold standard test. Results Two- part Bayesian model can detect the impact of covariates not only on disease status but also on test scores. Moreover, the areas under the ROC curve with different values of covariates could be calculated. The method was also trea- ted as the proof of statistical analysis for clinical diagnostic test. Conclnsion Two-part Bayesian model can be effectively used to solve the prob- lem of estimating ROC curves impacted by covariates when there is no gold standard test.

关 键 词：两部贝叶斯模型 ROC曲线 无金标准 协变量 

分 类 号：R195[医药卫生—卫生统计学]