Discriminant model |
Diagnostic class |
AUC (95% CI) |
Accuracy |
Sensitivity |
Specificity |
PPV |
NPV |
7 VOCs |
Recovered |
0.833 (0.731-0.935) |
69.12% |
68.75% |
69.23% |
40.74% |
87.80% |
|
Mild |
0.899 (0.826-0.972) |
70.59% |
91.67% |
66.07% |
36.67% |
97.37% |
|
Moderate |
0.865 (0.770-0.961) |
70.59% |
68.75% |
71.15% |
42.31% |
88.10% |
|
Severe |
0.869 (0.772-0.967) |
75.00% |
66.67% |
79.55% |
64.00% |
81.40% |
7 VOCs + PROs
|
Recovered |
0.952 (0.903-1) |
86.76% |
81.25% |
88.46% |
68.42% |
93.88% |
|
Mild |
0.930 (0.871-0.990) |
69.12% |
83.33% |
66.07% |
34.48% |
94.87% |
|
Moderate |
0.901 (0.820-0.983) |
70.59% |
62.50% |
73.08% |
41.67% |
86.36% |
|
Severe |
0.893 (0.799-0.987) |
83.82% |
83.33% |
84.09% |
74.07% |
90.24% |
PROs
|
Recovered |
0.845 (0.713-0.977) |
89.71% |
75.00% |
94.23% |
80.00% |
92.45% |
|
Mild |
0.532(0.355-0.709) |
33.82% |
83.33% |
23.21% |
18.87% |
86.67% |
|
Moderate |
0.644 (0.508-0.781) |
45.59% |
100.00% |
28.85% |
30.19% |
100.00% |
|
Severe |
0.638 (0.507-0.769) |
54.41% |
95.83% |
31.82% |
43.40% |
93.33% |