Objectives: The Centers for Disease Control and Prevention shifted the focus of surveillance paradigm for adult patients receiving mechanical ventilation, moving from the current standard of ventilator-associated pneumonia to broader complications. The surveillance definitions were designed to enable objective measures and efficient processes, so as to facilitate quality improvement initiatives and enhance standardized benchmark comparisons. We evaluated the surveillance definitions in term of their ability to predict clinical outcomes and ease of surveillance in a PICU.
Design: Retrospective cohort study.
Setting: A PICU at a university-affiliated children's hospital.
Patients: Eight hundred thirty-six patients receiving mechanical ventilation over 1-year period.
Interventions: None.
Measurements and main results: We applied the definition for ventilator-associated condition (i.e., a sustained increase in ventilator setting after a period of stable or decreasing support) to our database. Of total 606 patients, 14.5% had ventilator-associated condition (20.9/1,000 ventilator days) and 8.1% had an infection-related ventilator-associated condition (12.9/1,000 ventilator days). The patients with infection-related ventilator-associated condition were classified into probable pneumonia (55%), possible pneumonia (28.6%), and undetermined infection (16.3%). A large portion of patients with ventilator-associated condition (44%) had other noninfectious etiologies (e.g., atelectasis, pulmonary edema, and shock). Patients who developed ventilator-associated condition had significantly longer ventilatory, ICU, and hospital days compared with those who did not. The ventilator-associated condition group had increased hospital mortality compared with the non-ventilator-associated condition group (19.3% vs 6.9%; p=0.0007). Multivariate regression analysis identified ventilator-associated condition as one of the predictors of hospital mortality with an adjusted odds ratio of 2.14 (95% CI, 1.03-4.42). Risk factors for developing a ventilator-associated condition included immunocompromised status (odds ratio, 2.90; 95% CI, 1.57-5.33), tracheostomy dependence (odds ratio, 2.78; 95% CI, 1.40-5.51), and chronic respiratory disease (odds ratio, 1.85; 95% CI, 1.03-3.3).
Conclusions: The definitions for the various ventilator-associated conditions are good predictors of outcomes in children and adults and are amenable to automated surveillance. Based on the study findings, we suggest consideration for shifting the focus of surveillance for ventilator-associated events from only pneumonia to a broader range of complications.