Meta-Analysis of Predictive Test Accuracies
*Mun Sang Yue, Brown University
Keywords: meta-analysis, positive predictive value, negative predictive value, prevalence, ROC
Methodological developments in the meta-analysis of diagnostic test measures have been extensive and rich. In sharp contrast, methods for the meta-analysis of test predictive values---namely positive predictive values and negative predictive values---have not received much attention, despite the clinical relevance. As the use of diagnostic tests in precision medicine becomes increasingly common, methodological developments in the meta-analysis of test predictive values are needed. In this study, we propose a meta-analysis model that jointly synthesizes test accuracies and disease prevalence from different prospectively designed studies. The hierarchical summary ROC and summary predictive ROC curves generated from the model can be used for comparison of different tests. We applied the proposed method to a meta-analysis of prognostic capabilities of biomarkers for rapid rule-out of acute myocardial infarction and compared the prognostic capabilities of these biomarkers.