Table 2. A Multivariate Analysis Predicting 28-Day Mortality in Septic Patients from the Training Set.

	Model 1 (AUC: 0.7901)		Model 2 (AUC: 0.7901)		Model 3 (AUC: 0.8173)		Model 4 (AUC: 0.8173)
	OR (95% CI)	p value	OR (95% CI)	p value	OR (95% CI)	p value	OR (95% CI)	p value
Albumin	0.276 (0.178–0.427)	<0.001	0.276 (0.178–0.427)	<0.001	0.254 (0.161–0.403)	<0.001	0.254 (0.161–0.403)	<0.001
BE	0.899 (0.856–0.944)	<0.001	0.899 (0.856–0.944)	<0.001	0.901 (0.856–0.948)	<0.001	0.901 (0.856–0.948)	<0.001
RR					1.119 (1.057–1.183)	<0.001	1.119 (1.057–1.183)	<0.001

AUC, area under the receiver operating characteristic curve; OR, odds ratio; CI, confidence interval; BE, base excess; RR, respiratory rate; APACHE II, Acute Physiology and Chronic Health Evaluation II; SOFA, Sepsis Organ Failure Assessment; NEWS, National Early Warning Score.

Each scoring system was entered individually into the multivariate logistic regression analysis with clinical variables and laboratory variables in the training set. A total of four models were generated by a stepwise multivariate analysis method. Model 1 includes clinical variables, laboratory variables, and APACHE II; Model 2 includes clinical variables, laboratory variables, and NEWS; Model 3 includes clinical variables, laboratory variables, and SOFA; and Model 4 includes clinical variables and laboratory variables without a scoring system.