Table 8.

Comparison among the features used in exist state-of-the-art works and different combinations of the proposed features on ICU admission prediction on Site B dataset.

Features	AUC			Sensitivity (PPV = 70%)			K
	Mean	95% CI	p value	Mean	95% CI	p value
Img feature (Tang 2020)	0.770	(0.745, 0.796)	$p<0.001$	83.1%	(75.8%, 90.4%)	$p=0.009$	10
Img feature (Zhu 2020)	0.767	(0.752, 0.781)	$p<0.001$	83.8%	(82.2%, 85.5%)	$p<0.001$	39
DVB	0.671	(0.643, 0.700)	$p<0.001$	78.7%	(69.7%, 87.7%)	$p=0.007$	4
HLQ	0.791	(0.774, 0.809)	$p<0.001$	84.6%	(81.3%, 88.0%)	$p<0.001$	3
WLR	0.841	(0.827, 0.855)	$p=0.014$	94.9%	(93.4%, 96.3%)	–	55
WLR + HLQ	0.847	(0.833, 0.861)	–	92.6%	(89.5%, 95.7%)	p = 0.083	12
WLR+DVB	0.841	(0.828, 0.854)	p = 0.257	91.8%	(90.5%, 93.1%)	$p=0.012$	33
HLQ + DVB	0.796	(0.777, 0.815)	$p=0.001$	84.4%	(80.7%, 88.0%)	$p<0.001$	4
WLR + HLQ + DVB	0.844	(0.833, 0.855)	p = 0.310	92.6%	(90.0%, 95.1%)	$p=0.027$	12