Table 3 Sensitivity, specificity, and positive and negative predictive values according to different cutoff points in both, (a) derivation-internal validation and (b) external validation datasets.

From: Machine learning-based model for prediction of clinical deterioration in hospitalized patients by COVID 19

Risk groups

No patients

Derivation cohort

No (%) deteriorated in complementary risk groups

Sensitivity

Specificity

Positive predictive value

Negative predictive value

Score > 0.04

1424

1

0.12

0.26

0.99

1 (0.69%)

Score > 0.13

851

0.98

0.59

0.42

0.99

9 (1.26%)

Score > 0.23

530

0.90

0.83

0.62

0.96

38 (3.66%)

Score > 0.47

227

0.59

0.99

0.94

0.89

151 (11.26%)

Score > 0.82

31

0.08

1

1

0.78

334 (21.73%)

Risk groups

No patients

Validation cohort

No (%) deteriorated in complementary risk groups

Sensitivity

Specificity

Positive predictive value

Negative predictive value

Score > 0.04

862

0.99

0.12

0.22

0.99

1 (1.06%)

Score > 0.13

525

0.86

0.53

0.31

0.94

26 (6.03%)

Score > 0.23

313

0.65

0.75

0.40

0.90

67 (10.42%)

Score > 0.47

98

0.33

0.95

0.64

0.85

128 (14.92%)

Score > 0.82

13

0.06

1

0.92

0.81

179 (18.98%)

  1. Risk cut-off values were defined by the total point score for an individual, which represented low (< 2% mortality rate), intermediate (2–14.9%), or high risk (≥ 15%) groups, similar to commonly used pneumonia risk stratification scores.
Back to article page