Risk factor analysis and nomogram for predicting in-hospital mortality in ICU patients with sepsis and lung infection

Table 4 The AUROC and IDI of SOFA, SAPSII and Nomogram in training set and validation set

Predictive Model	AUROC		P^a value	IDI	P value
Training set	Nomogram	0.743 (0.713–0.773)
	SOFA	0.647 (0.616–0.684)	< 0.001	9.73% (7.48–11.98%)	< 0.001
	SAPSII	0.707 (0.668–0.741)	0.004	6.84% (4.60–9.08%)	< 0.001
Validation set	Nomogram	0.746 (0.699–0.790)
	SOFA	0.596 (0.543–0.654)	< 0.001	14.06% (10.71–17.41%)	< 0.001
	SAPSII	0.664 (0.613–0.715)	0.007	11.68% (8.06–15.30%)	< 0.001

AUROC area under the receiver operating characteristic curve; IDI integrated discrimination improvement
^aDelong’s test was used for testing the difference of AUROC between SOFA SAPS II scores and the nomogram model. In training set, result of AUROC between SOFA and the nomogram is Z = −5.0879 (p < 0.01), result of AUROC between SAPII and the nomogram is Z = 2.8677 (p = 0.004); in validation set,result of AUROC between SOFA and the nomogram is Z = −5.5984 (p < 0.01), result of AUROC between SAPII and the nomogram is Z = 2.7171 (p = 0.007)

ISSN: 1471-2466