Objective: Tracheal intubation in PICUs is associated with adverse tracheal intubation-associated events. Patient, provider, and practice factors have been associated with tracheal intubation-associated events; however, site-level variance and the association of site-level characteristics on tracheal intubation-associated event outcomes are unknown. We hypothesize that site-level variance exists in the prevalence of tracheal intubation-associated events and that site characteristics may affect outcomes.
Design: Prospective observational cohort study.
Setting: Fifteen PICUs in North America.
Subjects: Critically ill pediatric patients requiring tracheal intubation.
Interventions: None.
Measurement and main results: Tracheal intubation quality improvement data were collected in 15 PICUs from July 2010 to December 2011 using a National Emergency Airway Registry for Children with robust site-specific compliance. Tracheal intubation-associated events and severe tracheal intubation-associated events were explicitly defined a priori. We analyzed the association of site-level variance with tracheal intubation-associated events using univariate analysis and adjusted for previously identified patient- and provider-level risk factors. Analysis of 1,720 consecutive intubations revealed an overall prevalence of 20% tracheal intubation-associated events and 6.5% severe tracheal intubation-associated events, with considerable site variability ranging from 0% to 44% tracheal intubation-associated events and from 0% to 20% severe tracheal intubation-associated events. Larger PICU size (> 26 beds) was associated with fewer tracheal intubation-associated events (18% vs 23%, p = 0.006), but the presence of a fellowship program was not (20% vs 18%, p = 0.58). After adjusting for patient and provider characteristics, both PICU size and fellowship presence were not associated with tracheal intubation-associated events (p = 0.44 and p = 0.18, respectively). Presence of mixed ICU with cardiac surgery was independently associated with a higher prevalence of tracheal intubation-associated events (25% vs 15%; p < 0.001; adjusted odds ratio, 1.81; 95% CI, 1.29-2.53; p = 0.01). Substantial site-level variance was observed in medication use, which was not explained by patient characteristic differences.
Conclusions: Substantial site-level variance exists in tracheal intubation practice, tracheal intubation-associated events, and severe tracheal intubation-associated events. Neither PICU size nor fellowship training program explained site-level variance. Interventions to reduce tracheal intubation-associated event prevalence and severity will likely need to be contextualized to variability in individual ICUs patients, providers, and practice.