Table 5.

ROC analysis results

	AUC	Sensitivity	Specificity	Positive predictive value	Negative predictive value	Accuracy
Clinical/laboratory model
To identify discharged versus hospitalized/died patients	0.57	0.86	0.50	0.69	0.74	0.70
To identify discharged/stable patients versus critical/died patients	0.47	0.33	0.88	0.45	0.81	0.75
CT model
To identify discharged versus hospitalized/died patients	0.70	0.58	0.87	0.88	0.57	0.69
To identify discharged /stable patients versus critical/died patients	0.75	0.73	0.75	0.50	0.89	0.75
Clinical/laboratory and CT findings model
To identify discharged versus hospitalized/died patients	0.75	0.61	0.85	0.86	0.59	0.70
To identify discharged/stable patients versus critical/died patients	0.83	0.88	0.78	0.57	0.95	0.81