Skip to main content

Table 3 Comparison of MACPD per 100 beds according to same sensitivity

From: Early prediction of need for invasive mechanical ventilation in the neonatal intensive care unit using artificial intelligence and electronic health records: a clinical study

Models (≥ threshold)

Sensitivity

MACPD

NEWS ≥ 1

0.6798

1382

XGBoost(SpO2,FiO2) (≥ 0.2145)

0.6614

757

XGBoost (≥ 0.2569)

0.6798

374

Random forest (≥ 0.3943)

0.6798

406

Proposed (≥ 0.5588)

0.6798

226

NEWS ≥ 2

0.3248

383

XGBoost(SpO2,FiO2) (≥ 0.4885)

0.3276

125

XGBoost (≥ 0.7409)

0.3243

28

Random forest (≥ 0.5283)

0.3248

45

Proposed (≥ 0.8447)

0.3248

14

NEWS ≥ 3

0.1812

101

XGBoost(SpO2,FiO2) (≥ 0.6669)

0.1817

38

XGBoost (≥ 0.8494)

0.1826

8

Random forest (≥ 0.5836)

0.1812

15

Proposed (≥ 0.9126)

0.1812

6

NEWS ≥ 4

0.0838

39

XGBoost(SpO2,FiO2) (≥ 0.8099)

0.0828

5

XGBoost (≥ 0.9132)

0.0819

2

Random forest (≥ 0.6340)

0.0838

3

Proposed (≥ 0.9520)

0.0838

2

  1. Abbreviations: MACPD Mean alarm count per day, NEWS Newborn early warning score system, NICU Neonatal intensive care unit, XGBoost Extreme gradient boosting