Table 3.
Comparison of predictive efficiency of varied prognostic models.
| Models | 28 day prediction | ||||||
|---|---|---|---|---|---|---|---|
| AUROC | 95%CI | Z value | P value | Sensitivity | Specificity | Cut-off | |
| TATIM | 0.828 | 0.738–0.892 | 7.620 | <0.001 | 95.5 | 60.1 | −1.855 |
| CLIF-C ACLF | 0.797 | 0.706–0.870 | 6.441 | <0.001 | 100.0 | 50.6 | 36 |
| COSSH ACLF | 0.796 | 0.705–0.869 | 7.293 | <0.001 | 100.0 | 40.5 | 5 |
| SOFA | 0.729 | 0.635–0.811 | 4.091 | <0.001 | 58.33 | 78.3 | 7 |
| iMELD | 0.725 | 0.628–0.809 | 4.113 | <0.001 | 62.5 | 79.8 | 4 |
| MELD | 0.661 | 0.561–0.752 | 2.645 | <0.001 | 54.2 | 84.2 | 25 |
| Child-Pugh | 0.547 | 0.446–0.646 | 1.280 | 0.201 | 66.7 | 53.6 | 10 |
CLIF-C ACLF, ACLF model of European Association for the Study of Chronic Liver Failure; SOFA, chronic liver failure-sequential organ failure assessment; MELD, Model for End-stage Liver Disease; SOFA, sequential organ failure assessment; COSSH ACLF, ACLF model of Chinese Group on the Study of Severe Hepatitis B.