National variation in the use of tracheostomy in patients with congenital heart disease

Objectives: The postsurgical care of children with congenital heart disease may be complicated by the need for cardiorespiratory support, including tracheostomy. The variation of the use of tracheostomy across multiple pediatric cardiac surgical centers has not been defined. We describe multicenter variation in the use of tracheostomy in children undergoing congenital heart surgery. Design: We retrospectively analyzed a multicenter cohort. Setting: Pediatric Health Information Systems database retrospective cohort. Patients: Children less than 18 years who underwent both tracheostomy and cardiac surgery (1/04-6/14). Interventions: Univariate and multivariate statistics were performed, stratifying by high (≥ 75th percentile) and low (≤ 25th percentile) tracheostomy volume and adjusting for patient characteristics in multivariate models. Measurements and Main Results: Out of 123,510 hospitalizations involving cardiac surgery, 1,292 tracheostomies (1.2%) were performed (46 hospitals). The rate of tracheostomy placement ranged from 0.3% to 2.5% with no difference in the rate of tracheostomy placement between high and low tracheostomy use centers (p = 0.8). The median time to tracheostomy was 63 days (interquartile range, 36-100), and there was no difference between high- and low-tracheostomy centers. High-tracheostomy centers had $420,000 lower hospital charges than low-volume centers (p = 0.03). Tracheostomy day greater than the median (63 d), Risk Adjustment for Congenital Heart Surgery-1 score 6, and extracorporeal membrane oxygenation were significantly associated with adjusted increased odds of mortality. Later hospital day of tracheostomy was associated with a $13,000/d increase in total hospital charges (p < 0.001). Conclusions: Variation in the usage of tracheostomy in infants and children undergoing congenital heart surgery exists across the country. High-tracheostomy centers had lower hospital charges. Late tracheostomy placement, higher congenital heart disease surgical risk, and extracorporeal membrane oxygenation use are independent predictors of in-hospital mortality in this population.