Children with pneumonia presenting to the emergency department at Monroe Carell Jr. Children's Hospital at Vanderbilt or Children's Hospital of Pittsburgh will be potentially eligible for study. During intervention periods, providers caring for enrolled children will be presented with a detailed decision support strategy that emphasizes management in accordance with national guideline recommendations. The anticipated study duration is 18 months and, as this study does not include direct contact with enrolled subjects, there is no anticipated follow up.

Pneumonia is the most common serious infection in childhood. In the United States (US), pneumonia accounts for 1-4% of all emergency department (ED) visits in children (3-28 per 1,000 US children per year) and ranks among the top 3 reasons for pediatric hospitalization with >100,000 hospitalizations per year (15-22 per 100,000 US children per year)2-5. Pneumonia also accounts for more days of antibiotic use in US children's hospitals than any other condition.

Safely reducing inappropriate antibiotic use is critical to slow the progression of antimicrobial resistance, and childhood pneumonia is a key area where substantial improvements can be made. In the 2011 Pediatric Infectious Diseases Society/Infectious Diseases Society of America (PIDS/IDSA) pneumonia guideline, appropriate use of antibiotics was an important area of emphasis. For presumed bacterial pneumonia, recommendations emphasize the use of a single, narrow-spectrum antibiotic (i.e. amoxicillin or ampicillin). In children <6 years of age treated in the outpatient setting, the guideline recommended considering withholding antibiotics, recognizing that pneumonia in this population is most often caused by viruses. Both of these recommendations were graded as strong and supported by high-quality evidence. Nonetheless, in a large database study we conducted just prior to release of the 2011 guideline, use of broad-spectrum antibiotics was very common among children hospitalized with pneumonia, with substantial differences in antibiotic selection patterns among the various hospitals. In that study, use of narrow-spectrum ampicillin was rare (<5%). High rates of broad-spectrum antibiotic use were also noted in a study of children with pneumonia treated and released from US EDs, with <30% of children receiving narrow-spectrum therapy.

Study Aim: To test the hypothesis that electronic antibiotic decision support increases guideline-concordant antibiotic use compared with usual care in the emergency department.

The primary outcome is the proportion of children exclusively receiving guideline-concordant first line antibiotic therapy during the first 24 hours of care. Secondary outcomes include exclusive use of concordant antibiotic therapy for the entire episode, any use of concordant antibiotic therapy during the first 24 hours of care and for the entire episode, and emergency department revisits and hospitalizations within 72 hours and 7 days of the index discharge.

In this study, conducted at two experienced academic centers, the investigator will implement and evaluate an electronic health record based clinical decision support application to promote antibiotic use in concordance with the 2011 PIDS/IDSA guideline in a pragmatic, cluster randomized trial. Decisions regarding management, including antibiotic selection and site of care, will be at the discretion of the treating provider and will not be restricted or altered in any way. Thus, this study poses no greater than minimal risks to participants. Due to the nature of the research, waiver of informed consent has been provided, as it has been done in similar pragmatic studies at the institutions.

Inclusion Criteria:

• Six months to <18 years of age
• Radiographic evidence of pneumonia in ED
• Provider-confirmed diagnosis of pneumonia (via Clinical Decision Support [CDS])

Exclusion Criteria:

• Children with tracheostomy, cystic fibrosis, immunosuppression
• Inter-hospital transfers
• Hospitalization within preceding 7 days
• Previously enrolled within preceding 28 days
• Provider preference for any reason

18 months

Adverse event reporting will focus exclusively on ED revisits and/or inpatient hospital re-admissions within 3- or 7-days from index discharge and death within 30 days of index discharge. Adverse Events were monitored/assessed without regard to the specific Adverse Event Term.