Reducing Overuse of Antibiotics with Decision Support

Eliminating inappropriate antibiotic use in pediatric lower respiratory tract infections (LRTI) is the central focus of this research. LRTIs (pneumonia, bronchiolitis, and infection-related exacerbations of asthma) account for nearly one-third of all emergency department (ED) visits and 40% of all infection-related hospitalizations in US children. LRTIs also account for more antibiotic use in children's hospitals than any other condition, despite most LRTIs being viral in nature. Inappropriate antibiotics are associated with substantial adverse effects. Accordingly, national guidelines strongly discourage routine antibiotic use for bronchiolitis and acute asthma and argue for significantly reducing antibiotic exposure (initiation, spectrum, and duration) in pneumonia.

To address the problem of inappropriate antibiotic use, hospital-based antimicrobial stewardship programs (ASPs) are now common nationwide, and these programs have demonstrated effectiveness in some hospital settings. Unfortunately, traditional ASP approaches do not translate well to the fast-paced and unpredictable ED environment, and hospital-based ASP resources are finite and not always immediately available.

Clinical decision support (CDS) embedded within the electronic health record (EHR) is a strategy that could address the ED antibiotic stewardship gap. Informed by a deep understanding of the key facilitators and barriers to using CDS to support appropriate antibiotic use in ED and hospital settings, the investigators have developed two stewardship-focused CDS interventions for pediatric LRTI. The overarching goal of this research is to rigorously evaluate the implementation and effectiveness of these CDS tools, alone and in combination, against usual care only in a pragmatic randomized clinical trial at 3 U.S. children's hospitals.

This is a usual care-controlled superiority clinical trial platform designed to evaluate the effects of hospital-based CDS in the ED (CDS-ED) and after transitioning to the hospital setting (CDS-TR) on antibiotic prescribing and related clinical outcomes for child and adolescent LRTI encounters at 3 U.S. children's hospitals. The investigators hypothesize that both interventions will be superior to usual care and, among patients presenting in the ED and subsequently admitted to the hospital, the combined interventions (CDS-ED + CDS-TR) will be most effective overall. Randomization will occur sequentially in two stages corresponding to the ED CDS and Transitions CDS populations. The first stage of randomization will allocate qualifying ED encounters 1:1 to CDS-ED vs. usual care alone in the ED. The second stage of randomization will allocate participants requiring hospitalization (those discharged from the hospital are not eligible) 1:1 to CDS-TR vs. usual care at the time of admission. To minimize bias, the trial will be embedded within clinical care with minimal exclusions and disruption to usual care activities. Investigators will be blinded to study arm assignment, though blinding of treating clinicians is not possible due to the nature of the study. The trial will also evaluate process and implementation outcomes throughout the study period within the platform population. A formal interim analysis is not planned.