Bayesian两变量层次模型及其在诊断试验系统评价中的应用

目的探讨Bayesian两变量层次模型的构建及其在诊断试验系统评价中的应用。方法将Bayesian两变量层次模型应用于传统Pap细胞学涂片诊断子宫颈癌准确性评价的历史Meta分析资料,估计相关的效应指标敏感度和特异度及筛查研究比随访研究的相对可信度。结果与经典综合受试者工作特征曲线方法相比,Bayesian两变量层次模型估计得到三个层次的效应指标,其中综合敏感度和特异度均数及95％可信区间分别为0．64（0．56,0．72）和0．74（0．67,0．80）,预测敏感度和特异度均数及95％可信区间分别为0．61（0．12,0．96）和0．69（0．21,0．97）,筛查研究比随访研究的相对可信度估计为1．3（0.59,2.48）。结论采用Bayesian两变量层次模型进行诊断试验Meta分析,更加灵活、有效,易于实现和解释,值得推广应用。

Application of Bayesian Bivariate Hierarchical Model in Meta Analysis of Diagnostic Test

Objective To explore the application of Bayesian bivariate hierarchical model in meta analysis of diagnostic test. Methods Bayesian bivariate hierarchical model was utilized in the meta analysis of accuracy of conventional Pap cytology smear test for cervical cancer screening to estimate sensitivity and speciality of effect index and relative credibility of screening and follow-up studies. Result Three levels effect index were obtained, the means and 95% credible intervals of pooled sensitivity and speciality were 0.64 (0.56,0.72) and 0.74 (0.67,0.80), of the predicted sensitivity and speciality were 0.61 (0.12,0.96) and 0.69(0.21 ,0.97), respectively. Estimation of relative credibility of screening over follow-up studies was 1.3 (0.59,2.48). Conclusion Bayesian bivariate hierarchical model is a flexible and effective method that is easy to realize and interpret for meta analysis of diagnostic test and worthy of being utilized and popularized.