Table 2.

Performance metrics for the machine learning algorithm and all comparator scores for mortality prediction on the total study population.

Statistics	MLAa	MEWSb	APACHEc	SAPS-IId
AUROCe	0.780	0.580	0.700	0.660
Sensitivity	0.751	0.523	0.678	0.674
Specificity	0.656	0.577	0.596	0.511
DORf	5.739	1.499	3.106	2.157
LR+g	2.181	1.238	1.678	1.378
LR–h	0.380	0.826	0.540	0.639

^aMLA: machine learning algorithm.

^bMEWS: Modified Early Warning Score.

^cAPACHE: Acute Physiologic Assessment and Chronic Health Evaluation.

^dSAPS II: Simplified Acute Physiology Score II.

^eAUROC: area under the receiver operating characteristic.

^fDOR: diagnostic odds ratio.

^gLR+: positive likelihood ratio.

^hLR–: negative likelihood ratio.