ROC curves plotted for the Predictive Power of Scoring Systems in mortality. (A) For the performance of pGCS in predicting mortality, the AUC was calculated as 0.943, sensitivity was 100%, and specificity was 84.4% at the cutoff of 10 (p < 0.001). For the performance of PTS in predicting mortality, the AUC was calculated as 0.951, sensitivity was 92.3%, and specificity was 87.2% at the cutoff of 3.5 (p < 0.001). (B) For the performance of PEWS in predicting mortality, the AUC was calculated as 0.953, sensitivity was 92.3%, and specificity was 93.6% at the cutoff of 6.5 (p < 0.001). For the performance of PRISM-3 in predicting mortality, the AUC was calculated as 0.999, sensitivity was 100%, and specificity was 99.1% at the cutoff of 21 (p < 0.001).