Table 6 Demographic characteristics of the populations of ICU patients admitted during the preintervention and postintervention periods, of the patients with clinically suspected VAP and the effects of the intervention on diagnosis of VAP and antibiotic use
Preintervention periodPostintervention periodDifference (95% CI)p Value
Entire population admitted to ICU
Total number of patients10591075
% male58%59%−1% (−5% to 3%)0.43*
Mean age (years)55.954.41.5 (0 to 3)0.06†
Median (IQR) APACHE II score17 (12–23)17 (11–23)00.91‡
Median (IQR) length of stay (days)1.9 (0.9–4.9)2.1 (0.9–4.850.20.78‡
ICU mortality19%20%−0.7% (−4% to 2.6%)0.6*
Patients “at risk” (ie, length of stay ⩾48 h)643667
Number of at-risk “ventilator days”37713777
Population with clinically suspected VAP
Number with clinically suspected VAP110 (17%)94 (14%)−3% (−6% to 1%)0.16*
Incidence of suspected VAP (per 1000 ventilator days)29.124.9
Median (IQR) APACHE II score23 (18–26.5)20 (16–24)30.02‡
Number of suspected VAPs undergoing bronchoscopic diagnosis41 (37%)53(58%)19% (5 to 32%)0.004*
Cases of microbiologically confirmed VAP68 (62%)36 (40%)−22% (−36% to −10%)0.0012*
% Gram-negative organisms51%62%11% (−37% to 15%)0.35*
Incidence of confirmed VAP (per 1000 ventilator days)189
Mean “antibiotic days”9.17.2−1.8 (−3.8 to 0.4)0.08†
Mortality37.5%17%−20.5% (−22% to −19%)0.003*
1Median (IQR) duration of ventilation16 (10–30)19 (10–27)30.6
  • *By z test for proportions.

  • †By t test.

  • ‡By Mann–Whitney U test.

  • APACHE II, Acute Physiology and Chronic Health Evaluation II; BAL, bronchoalveolar lavage; ETA, endotracheal aspirate; ICU, Intensive Care Unit; IQR, interquartile range; VAP, ventilator-associated pneumonia.