Fig. 1From: Predicting patient-reported outcomes following lumbar spine surgery: development and external validation of multivariable prediction modelsDiscrimination ability of the (a) logistic regression and (b) random forest classification models when fitted to the validation data for MCID in COMI, back pain, and leg pain. Plots illustrate Receiver-Operating Characteristic (ROC) curves with an optimal probability threshold (black point on the ROC curve; specificity and sensitivity indicated in brackets). Area Under the Curve (AUC) is reported for each ROC with 95% confidence intervalBack to article page