Development and validation of a novel blending machine learning model for hospital mortality prediction in ICU patients with Sepsis

Table 2 Baseline characteristics between training set and test set

	Train set		Test set		P-value
Data source	eICU-CRD		MIMIC-III
Number of hospitals	198		1
Number of patients	12,558		12,095
Unit type, n (%)	Med-Surg ICU	8255 (65.7)	MICU	4938 (40.8)
	MICU	1479 (11.8)	SICU	1839 (15.2)
	CCU	927 (7.4)	CCU	1328 (11.0)
	CCU-CTICU	775 (6.2)	CSRU	2459 (20.3)
	SICU	541 (4.3)	TSICU	1531 (12.7)
	NICU	302 (2.4)
	CSICU	135 (1.0)
	CTICU	144 (1.1)
SOFA score (mean (SD))	5.3 (2.8)		5.5 (3.1)		< 0.001
Length of ICU stay (hours, median [IQR])	70.0 [43.0, 127.0]		74.0 [43.9, 155.8]		< 0.001
Hospital mortality (%)	17.9		17.9		0.999

Abbreviations: SD standard deviation, IQR interquartile range, CCU cardiac care unit, CSICU cardiac surgical intensive care unit, CSRU cardiac surgery recovery unit, CTICU cardiothoracic intensive care unit, MICU medical intensive care unit, NICU neuro-intensive care unit, SICU surgical intensive care unit, TSICU trauma surgical intensive care unit

ISSN: 1756-0381