Medical Student Hackensack Meridian School of Medicine Hoboken, New Jersey, United States
Background: In the 2014 Clinical Practice Guideline: The Diagnosis, Management, and Prevention of Bronchiolitis, the American Academy of Pediatrics strengthened its recommendation against administering albuterol. Since then, many quality improvement (QI) efforts have demonstrated success in reducing albuterol use and improving evidence-based care delivery for hospitalized children with bronchiolitis. However, little is known about how other nebulized therapies are used in the wake of efforts to reduce albuterol use. Objective: To examine if efforts to reduce albuterol use are associated with reciprocal changes in use of similarly unproven bronchiolitis treatment therapies, including racemic epinephrine, normal saline, and hypertonic saline. We hypothesize that decreasing albuterol use will result in increased use of unproven therapies, and this increase will vary by institution. Design/Methods: This is a secondary analysis of an ongoing, multicenter QI project across an integrated health network. The network includes two academic children’s hospitals and a community hospital with a pediatric inpatient service. The QI project implemented system changes (baseline: Nov 21-Oct 22; intervention: Jan 23-Dec 23) aimed at reducing albuterol use in children 1 to 23 months-old admitted with bronchiolitis and no disqualifying conditions (e.g., prematurity, congenital heart disease, immunodeficiency). For this secondary analysis, the dependent variable is the count of patients receiving at least one dose of nebulized therapies (either racemic epinephrine, normal saline, or hypertonic saline) during their admission. Independent variables of interest include: institution, study phase, age, and season. Descriptive statistics will summarize the data and identify associations through univariate and bivariate analyses. A poisson regression model will assess use of individual therapies before and after the implemented system changes. Variables of interest and significant confounders (alpha=0.05) identified in bivariate analyses will be added to the model, and incidence rates will be reported.
