Skip to main content
. 2015 Jun 21;15:241. doi: 10.1186/s12913-015-0917-x

Table 2.

All-cause mortality of ICU patients with S. aureus or P. aeruginosa pneumonia versus controls

Cohorts of interest Comparisons
S. aureus pneumonia vs No pneumonia P. aeruginosa pneumonia vs No pneumonia
No pneumonia (n = 201,394) S. aureus pneumonia (n = 2275) P. aeruginosa pneumonia (n = 1007) Odds Ratio/Differencea (95 % CI) P Valuea Odds Ratio/Differenceb (95 % CI) P Valueb
Index hospitalization mortality
No. (%) of patients 6,188 (3.1) 363 (16.0) 203 (20.2) 5.99 (5.34, 6.72) <.001 7.97 (6.81, 9.31) <.001
Post-index mortality
No. (%) of patients 5,609 (2.8) 227 (10.0) 117 (11.6) 3.87 (3.37, 4.45) <.001 4.59 (3.78, 5.57) <.001
Mean (SD) time to deathc 32.8 (27.1) 25.5 (25.8) 28.1 (28.4) -7.30 (-10.51, -3.63) <.001 -4.69 (-9.44, 1.02) .10
Survival rate: 3.75 (3.28, 4.28) <.001 4.40 (3.66, 5.28) <.001
30 days (%) 98.5 93.4 92.9
90 days (%) 97.2 90.0 88.4
Index + post-index mortality
No. (%) of patients 11,797 (5.9) 590 (25.9) 320 (31.8) 5.63 (5.11, 6.19) <.001 7.49 (6.55, 8.56) <.001
Mean (SD) time to deathc 24.9 (28.3) 46.4 (47.6) 63.5 (59.3) 21.53 (17.39, 26.07) <.001 38.57 (31.08, 47.07) <.001

CI confidence interval, ICU intensive care unit

aOdds ratio (OR) from Chi-square test is used for categorical variables, negative binomial models are used to test for differences between means for count variables (number of events and length of stay). Statistical comparisons are comparing Cohort 3 (pneumonia due to S. aureus) with Cohort 1 (no pneumonia, reference group); ie, Difference = mean (Cohort 3) − mean (Cohort 1) and OR = Odds (Cohort 3)/Odds (Cohort 1)

bOdds ratio (OR) from Chi-square test is used for categorical variables, negative binomial models are used to test for differences between means for count variables (number of events and length of stay). Statistical comparisons are comparing Cohort 4 (pneumonia due to P. aeruginosa) with Cohort 1 (no pneumonia, reference group); i.e., Difference = mean (Cohort 4) − mean (Cohort 1) and OR = Odds (Cohort 4)/Odds (Cohort 1)

cIncluding only patients with mortality