Table 2 Discrimination performance of sudden cardiac arrest.

From: Real-time machine learning model to predict in-hospital cardiac arrest using heart rate variability in ICU

 

AUROC

AUPRC

Sensitivity

Specificity

Precision

Accuracy

F1-score

Primary outcome

 Within 0.5–24 h

0.881 (0.875–0.887)

0.104 (0.093–0.116)

0.817 (0.800– 0.834)

0.800 (0.798–0.802)

0.053 (0.051–0.056)

0.809 (0.800–0.817)

0.100 (0.095–0.104)

Secondary outcomes

 Within 0.5–18 h

0.881 (0.875–0.887)

0.105 (0.093–0.117)

0.814 (0.796–0.832)

0.800 (0.798–0.802)

0.052 (0.049–0.054)

0.807 (0.798–0.816)

0.097 (0.093–0.102)

 Within 0.5–12 h

0.885 (0.878–0.891)

0.098 (0.086–0.111)

0.827 (0.806–0.846)

0.800 (0.798–0.802)

0.044 (0.042–0.047)

0.813 (0.803–0.823)

0.084 (0.080–0.089)

 Within 0.5–6 h

0.883 (0.874–0.892)

0.070 (0.059–0.083)

0.840 (0.817–0.862)

0.800 (0.798–0.802)

0.030 (0.028–0.032)

0.820 (0.809–0.831)

0.058 (0.055–0.062)

 Within 0.5–3 h

0.879 (0.865–0.892)

0.038 (0.029–0.049)

0.850 (0.817–0.881)

0.800 (0.798–0.802)

0.016 (0.014–0.017)

0.825 (0.809–0.841)

0.031 (0.028–0.034)

 Within 0.5–1 h

0.858 (0.827–0.885)

0.004 (0.003–0.006)

0.823 (0.755–0.888)

0.800 (0.798–0.802)

0.004 (0.003–0.004)

0.812 (0.778–0.844)

0.007 (0.006–0.009)

  1. Data are presented as mean with 95% confidence interval. AUROC area under the receiver operating characteristic curve, AUPRC area under the precision-recall curve.
Back to article page