Skip to main content
. 2022 Jan 21;13:811952. doi: 10.3389/fimmu.2022.811952

Table 2.

The accuracy of the model to distinguish severity groups 2b and 3 from severity groups 1 and 2a in the validation set.

A. Clinical data alone
Day True Severity Estimated S1, 2a (n) Estimate S2b, 3 (n) Error Rate (%) Accuracy
day 4-7 S1, 2a (n) 22 0 0.00 0.93
S2b, 3 (n) 2 3 0.40
day 5-8 S1, 2a (n) 25 2 0.07 0.80
S2b, 3 (n) 5 3 0.63
day 6-9 S1, 2a (n) 26 1 0.04 0.92
S2b, 3 (n) 2 10 0.17
day 7-10 S1, 2a (n) 26 3 0.10 0.77
S2b, 3 (n) 7 8 0.47
day 8-11 S1, 2a (n) 22 8 0.27 0.77
S2b, 3 (n) 4 18 0.18
day 9-12 S1, 2a (n) 25 4 0.14 0.84
S2b, 3 (n) 5 22 0.23
B. Clinical data + antibody data
Day True Severity Estimated S1 (n) Estimate S2a, 2b, 3 (n) Error Rate (%) Accuracy
day 4-7 S1, 2a (n) 22 0 0.00 0.96
S2b, 3 (n) 1 4 0.20
day 5-8 S1, 2a (n) 26 1 0.04 0.91
S2b, 3 (n) 2 6 0.25
day 6-9 S1, 2a (n) 25 2 0.07 0.95
S2b, 3 (n) 0 12 0.00
day 7-10 S1, 2a (n) 29 0 0.00 0.89
S2b, 3 (n) 5 10 0.33
day 8-11 S1, 2a (n) 25 5 0.17 0.88
S2b, 3 (n) 1 21 0.05
day 9-12 S1, 2a (n) 26 3 0.10 0.84
S2b, 3 (n) 6 21 0.22