Table 3.

Models of Multiple Logistic Regression Employing Different Independent Variables of Severity Outcome

Models	Variables	Categories	DF	Estimate	SE	χ2	P	OR (95% CI)
Model 1	Intercept	Hospital stay ≥20 days	1	−2.143	0.305	49.277	<0.001
		Hospital stay 10–19 days	1	3.103	0.457	46.022	<0.001
	Sleep status (Recommended as reference)	Maybe appropriate	1	1.907	0.585	10.620	0.001*	6.729 (2.138–21.181)
		Lack of sleep	1	2.153	0.768	7.871	0.005*	8.612 (1.913–38.760)
Model 2	Intercept	Hospital stay ≥20 days	1	3.621	1.643	4.859	0.028
		Hospital stay 10– 19 days	1	8.773	1.834	22.889	<0.001
	Average daily sleep time (h)		1	−0.771	0.243	10.041	0.002*	0.463 (0.287–0.745)

Notes: Significant level is 0.05 and significant P-values are shown with an asterisk. Model 1: took clinical severity as the dependent variable and Sleep status (Categorical variable) as one of independent variables; Model 2: took clinical severity as the dependent variable and Average daily sleep time (hours) as one of the independent variables. Score Test for the Proportional Odds Assumption: Model 1: χ²=0.1570, DF=2, P=0.9245; Model 2: χ²=0.0012, DF=1, P=0.9726, which suggested that these two models satisfy the assumption.

Abbreviations: SE, standard error; OR, odds ratio; P, P-value of multiple logistic regression.