Table 4. Results from the logistic regression models – SA diagnosis vs. Non-SA diagnosis.

Multivariable regression for various outcomes. Present diagnosis of sleep apnea is the effect variable (CI = confidence interval).

Outcome	Crude Odds Ratio (Corrected 95% CI) a,b,c	Adjusted Odds Ratio (Corrected 95% CI) a,c	Significant interaction term with SA (p-value d)	C-statistic
Combined Complications	1.53 (1.47, 1.60)	1.47 (1.39, 1.55)	None	0.69
Pulmonary Complications	2.19 (2.00, 2.40)	1.86 (1.65, 2.09)	None	0.69
Cardiac Complications	1.56 (1.48, 1.65)	1.59 (1.48, 1.71)	None	0.74
Mortality	1.27 (0.82, 1.98)	1.27 (0.74, 2.19)	None	0.74
Mechanical Ventilation	11.94 (10.64, 13.40)	See table 5	Complicated Hypertension (p<0.0001) COPD (p<0.0001)	0.83
Non-Invasive Ventilation	37.13 (30.64, 45.00)	See table 5	Complicated Hypertension (p<0.0001), COPD (p<0.0001)	0.88
Blood Product Transfusion	0.71 (0.68, 0.74)	0.88 (0.83, 0.93)	None	0.67
ICU Utilization	2.06 (1.92, 2.21)	See table 5	Gender (p=0.0002), Year (p=0.0007)	0.73
Telemetry / Stepdown Unit Utilization	1.76 (1.66, 1.86)	See table 5	Year (p=0.0008), Complicated Hypertension(p<0.0001), Obesity (p<0.0001)	0.68
Length of Stay > 75th percentile	1.09 (1.05, 1.13)	See table 5	Obesity (p=0.0006)	0.65
Utilization of economic resources > 75th percentile	1.22 (1.18, 1.26)	1.14 (1.09, 1.19)	None	0.60

Reference = no sleep apnea diagnosis.

^aAll 95% CIs and the associated p-values were both Bonferroni corrected for multiple comparisons

^bAll corrected p-values were <.0001 except for mortality (p>0.99).

^cAll corrected p-values were <.0001 except for mortality (p>0.99).

^dP-values were raw p-values from multivariate regressions, and compared with threshold 0.05/11=0.004 to determine the statistical significance.