Heterogeneity in Bronchodilator Use for Bronchiolitis Across Pediatric Emergency Departments: A Multicenter PEDSnet Study

Background Despite 2014 guidelines discouraging routine bronchodilator use for bronchiolitis, substantial practice variation persists across pediatric emergency departments (EDs). Objective To examine demographic, clinical, and institutional factors associated with bronchodilator use and to develop a model identifying predictors of its administration. Methods We conducted a multicenter, retrospective study of children 3 to 24 months old with bronchiolitis seen in EDs from nine PEDSnet institutions (2019). Exclusions included prematurity, comorbidities, and prior asthma or reactive airway disease. Associations between patient characteristics and bronchodilator use were examined. A random forest model was trained and tested to identify correlates of bronchodilator administration. Results Among 14,213 visits, bronchodilators were administered in 29%. Use varied by site, age, steroid administration, respiratory rate, and prior diagnoses. The model achieved 85% accuracy (p < 0.001), 25% sensitivity, and 98% specificity. Key correlates included steroid use, geographic site, and age. Conclusion Bronchodilator administration for bronchiolitis remains highly variable and is associated with demographic, clinical, and geographic factors. While our model demonstrates the ability to identify patients unlikely to receive bronchodilators, its low sensitivity limits clinical utility. Findings underscore the need for future prospective studies exploring phenotype-driven bronchiolitis care, while cautioning against overinterpreting associations as evidence of therapeutic benefit.