Properties of the summary receiver operating characteristic (SROC) curve for diagnostic test data.

摘要: The summary receiver operating characteristic (SROC) curve has been recommended to represent the performance of a diagnostic test, based on data from meta-analysis. However, little is known about basic properties SROC or its estimate. In this paper, position characterized in terms overall odds ratio and magnitude inter-study heterogeneity ratio. area under (AUC) an index Q(*) are discussed as potentially useful summaries curve. It shown that AUC maximized when study ratios homogeneous, it quite robust heterogeneity. An upper bound derived for exact analytic expression homogeneous situation, lower limit case Q(*), defined by point where sensitivity equals specificity: invariant standard error studies, be reasonable approximation with heterogeneous studies. expressions easily computed case, avoid need numerical integration more general case. SE(AUC) SE(Q(*)) found numerically close, being larger if very large. methods illustrated using Pap smear screening test cervical cancer, three tests diagnosis metastases cancer patients.