Erratum to “ROC curves for clinical prediction models part 1. ROC plots showed no added value above the AUC when evaluating the performance of clinical prediction models” [J Clin Epidemiol. 126C(2020):207-16

Abstract

The publisher regrets that the incorrect figure 1 and figure 4 were published. The article has been corrected. The publisher would like to apologize for any inconvenience caused. Fig. 4Effect of increasing sample size (n) on the variability and shape of the ROC curve. We considered a normally distributed risk prediction model with a true AUC of 0.74 for predicting the presence of an event with a true event rate of 22%. For each of four sample sizes, 100 data sets were simulated and their corresponding ROC curves plotted. At lower sample sizes, ROC curves are more rough and variable than at high sample sizes.View Large Image Figure ViewerDownload Hi-res image Download (PPT) ROC curves for clinical prediction models part 1. ROC plots showed no added value above the AUC when evaluating the performance of clinical prediction modelsJournal of Clinical EpidemiologyVol. 126PreviewReceiver operating characteristic (ROC) curves show how well a risk prediction model discriminates between patients with and without a condition. We aim to investigate how ROC curves are presented in the literature and discuss and illustrate their potential limitations. Full-Text PDF