Table 2 Performance metrics of the models in predicting 30-day mortality and ICU admission.

From: Multimodal deep learning for COVID-19 prognosis prediction in the emergency department: a bi-centric study

 

Specificity

Recall

Precision

NPV

F1-score

MCC

AUC

30-Day mortality

 TensorFlow tabular model

0.76 ± 0.03

0.76 ± 0.04

0.49 ± 0.03

0.93 ± 0.01

0.58 ± 0.04

0.48 ± 0.05

0.88 ± 0.02

 Fastai tabular model

0.75 ± 0.06

0.77 ± 0.09

0.43 ± 0.04

0.93 ± 0.02

0.55 ± 0.03

0.43 ± 0.04

0.84 ± 0.02

 XGBoost model

0.75 ± 0.06

0.77 ± 0.09

0.43 ± 0.04

0.93 ± 0.02

0.55 ± 0.03

0.43 ± 0.04

0.84 ± 0.02

 TensorFlow tabular-textual model

0.82 ± 0.08

0.77 ± 0.13

0.53 ± 0.12

0.94 ± 0.03

0.62 ± 0.10

0.52 ± 0.13

0.87 ± 0.06

 Fastai tabular-textual model

0.78 ± 0.07

0.74 ± 0.08

0.46 ± 0.06

0.93 ± 0.02

0.56 ± 0.04

0.44 ± 0.04

0.84 ± 0.02

ICU admission

 TensorFlow tabular model

0.72 ± 0.05

0.76 ± 0.10

0.27 ± 0.03

0.96 ± 0.01

0.40 ± 0.03

0.33 ± 0.05*

0.83 ± 0.04

 Fastai tabular model

0.71 ± 0.06

0.70 ± 0.11

0.25 ± 0.03

0.95 ± 0.01

0.37 ± 0.04

0.29 ± 0.05

0.79 ± 0.04

 XGBoost model

0.71 ± 0.06

0.70 ± 0.10

0.25 ± 0.03

0.95 ± 0.01

0.37 ± 0.04

0.29 ± 0.05

0.79 ± 0.04

 TensorFlow tabular-textual model

0.92 ± 0.03

0.53 ± 0.15

0.48 ± 0.01

0.94 ± 0.02

0.49 ± 0.10

0.43 ± 0.11*

0.80 ± 0.07

 Fastai tabular-textual model

0.72 ± 0.08

0.67 ± 0.11

0.25 ± 0.03

0.94 ± 0.01

0.36 ± 0.03

0.28 ± 0.05

0.79 ± 0.04

  1. Results are presented as means and standard deviations of 10 iterations. NPV, negative predictive value; MCC, Matthew’s correlation coefficient; AUC, area under the curve. *p < 0.05 according to a paired t-test.
Back to article page