Table 2 Results of Deep Learning Survival Cox analyses on the three independent external validation cohorts.
From: Early triage of critically ill COVID-19 patients using deep learning
Cohort | Wuhan | Hubei | Guangdong | |||
|---|---|---|---|---|---|---|
Ex3 | All cases | Ex3 | All cases | Ex3 | All cases | |
No. of patients (critically) | 801 (84) | 940 (94) | 305 (8) | 380 (9) | 73 (3) | 73 (3) |
AUC (95% CI) | 0.893 (0.867–0.919) | 0.881 (0.854–0.905) | 0.888 (0.732–0.984) | 0.819 (0.632–0.978) | 0.967 (0.905–1.000) | 0.967 (0.905–1.000) |
C-index (95% CI) | 0.890 (0.865–0.915) | 0.878 (0.852–0.903) | 0.852 (0.672–0.973) | 0.769 (0.556–0.966) | 0.967 (0.906–1.000) | 0.967 (0.906–1.000) |
HR&MR recall (95% CI) | 1.000 (1.000–1.000) | 1.000 (1.000–1.000) | 0.875 (0.625–1.000) | 0.778 (0.500–1.000) | 1.000 (1.000–1.000) | 1.000 (1.000–1.000) |
HR recall (95% CI) | 0.833 (0.768–0.900) | 0.809 (0.736–0.878) | 0.500 (0.167–0.800) | 0.444 (0.167–0.750) | 0.667 (0.000–1.000) | 0.667 (0.000–1.000) |
