Table 1 Univariate analysis of the selected features for COVID-19 patients in the training cohort.

From: Early triage of critically ill COVID-19 patients using deep learning

Critical illness

 

Total (n = 1590)

No (n = 1459)

Yes (n = 131)

Hazard ratio (95% CI)

p-value

AUC (95% CI)

C-index (95% CI)

Age

48.9 ± 15.7

47.8 ± 15.2

61.6 ± 14.8

1.059 (1.046–1.071)

<0.001

0.755 (0.695–0.812)

0.732 (0.674–0.790)

Dyspnea

331/1394 (23.7)

257/1275 (20.2)

74/119 (62.2)

5.759 (3.973–8.346)

<0.001

0.665 (0.590–0.745)

0.659 (0.584–0.739)

Cancer history

18/1590 (1.1)

11/1459 (0.8)

7/131 (5.3)

5.927 (2.766–12.7)

<0.001

0.498 (0.495–0.500)

0.498 (0.495–0.500)

COPD

24/1590 (1.5)

12/1459 (0.8)

12/131 (9.2)

7.471 (4.124–13.53)

<0.001

0.532 (0.495–0.580)

0.516 (0.495–0.549)

No. of comorbidity

   

1.67 (1.506–1.851)

<0.001

0.697 (0.613–0.789)

0.682 (0.597–0.772)

 0

1191/1590 (74.9)

1137/1459 (77.9)

54/131 (41.2)

    

 1

269/1590 (16.9)

229/1459 (15.7)

40/131 (30.5)

    

 2

88/1590 (5.5)

68/1459 (4.7)

20/131 (15.3)

    

 3

34/1590 (2.1)

20/1459 (1.4)

14/131 (10.7)

    

 4

5/1590 (0.3)

4/1459 (0.3)

1/131 (0.8)

    

 5

3/1590 (0.2)

1/1459 (0.1)

2/131 (1.5)

    

X-ray abnormality

243/1590 (15.3)

184/1459 (12.6)

59/131 (45)

5.315 (3.765–7.504)

<0.001

0.600 (0.524–0.681)

0.614 (0.535–0.696)

Neutrophil/lymphocytes

5.1 ± 5.6

4.4 ± 3.8

12.7 ± 12.4

1.061 (1.052–1.071)

<0.001

0.861 (0.803–0.918)

0.857 (0.795–0.914)

Lactate dehydrogenase

314.3 ± 693.7

273.6 ± 135.2

723.6 ± 2239.5

1 (1–1)

<0.001

0.810 (0.745–0.868)

0.787 (0.727–0.846)

Direct bilirubin

4 ± 2.7

3.7 ± 2.3

6.5 ± 4.1

1.212 (1.165–1.261)

<0.001

0.674 (0.567–0.782)

0.662 (0.551–0.773)

Creatine kinase

135.5 ± 246.7

123 ± 125.3

258.9 ± 702.8

1.001 (1–1.001)

<0.001

0.557 (0.449–0.665)

0.554 (0.441–0.666)

  1. Data are mean ± SD, n/N (%), where N is the total number of patients with available data. P-values are calculated by log rank test.