Fig 1.
Diagnostic diagram for ventilator-associated pneumonia, based on the definitions of the former Centers for Disease Control and Prevention/National Nosocomial Infections Surveillance system, currently the National Healthcare Safety Network.
Fig 2.
Average yearly incidence density, per 1000 ventilation days, of ventilator-associated pneumonia, per hospital.
Hospitals with significant reduction are indicated with an asterisk (*).
Table 1.
Patient and ventilation characteristics and hazard ratios, including 95% confidence intervals and adjusted for hospital, of time-independent covariates.
Table 2.
Ventilation day characteristics and hazard ratios for univariate models of time-dependent covariates.
Fig 3.
Daily hazards of time-dependent covariates, in univariate models.
The curve shows the estimated risk attributed to exposures on each day prior to the last day of follow-up (i.e., the event date or the censoring date) and the grey ribbon shows the 95% confidence interval. A value of one indicates no effect of the exposure at that time. At times where the grey ribbon includes one, the effect is considered to be statistically insignificant.
Table 3.
Hazard ratios of patient and ventilation characteristics–results of multivariate analysis.
Fig 4.
Daily hazard ratios attributed to past values of the covariates assessed with a weighted cumulative effects approach in the multivariate Cox regression model: (A) ivSDD, (B) ivAB, (C) Jet nebulizer (compared to no inhalation therapy) and metered dose inhalers (MDI) (compared to no inhalation therapy). The curve shows the estimated risk attributed to exposures on each day prior to the last day of follow-up (i.e., the event date or the censoring date) and the grey ribbon shows the 95% confidence interval. A value of one indicates no effect of the exposure at that time. At times where the grey ribbon includes one, the effect is considered to be statistically insignificant.