Variability in Observer Performance Studies

摘要: Rationale and Objectives The aim of the study is to assess variance components in observer performance studies possible impact on results conclusions. Materials Methods Two previously performed retrospective receiver operating characteristic–type evaluate seven radiologists detecting interstitial disease conventional posteroanterior chest films nine a high-resolution workstation were reanalyzed by using Beiden, Wagner, Campbell nine-component model estimate different components. We estimated case-, reader-, mode-related for group as whole after excluding (round robin) each reader. Overall was evaluated, effect individual readers overall conclusions assessed. Results reanalysis agreed with original one that, group, significantly better when (P Conclusion Estimates can shed light appropriateness design, well sensitivity inclusion (or exclusion) observers.