Is there an improvement in performance when radiographs are re-reported at 24 hours?
GR Tudor and DB Finlay Department of Radiology, Leicester Royal Infirmary, NHS Trust, UK. This study was designed to assess the accuracy and intraobserver variability when plain radiographs are re-reported after 24 h, in order to determine if there was a change in observer performance. Five consultant radiologists reported 50 plain radiographs on patients in whom the diagnosis had been established. The radiographs were reviewed by each radiologist after an interval of 24 h. The same clinical information was available on both occasions. 32 of the radiographs showed an abnormality and 18 were normal. The accuracy of the report in relation to the established diagnosis was assessed, intraobserver agreement being calculated using kappa (kappa) statistics. The kappa-value for intraobserver agreement was in the range 0.33-0.88. Accuracy showed a range of 78-90% on the first viewing and 80-92% on the subsequent viewing. ROC curves were produced for each viewing for the group as a whole. The ar
