Resident Physician University of Minnesota University of Minnesota Masonic Children's Hospital Minneapolis, Minnesota, United States
Background: The 2014 AAP Guidelines on Bronchiolitis recommend against the use of chest radiography (CXR) as it results in overuse of antibiotics without improving outcomes. During a secondary analysis of a Quality Improvement (QI) initiative to decrease CXRs for bronchiolitis at a community Emergency Department (ED), researchers discovered discordant interpretations between general radiologists (GR) and pediatric radiologists (PR).
Objective: Evaluate discordant CXR reads at a community ED by comparing the interpretation provided by GR versus PR and introduce the clinical impact of misreads on patients.
Design/Methods: We analyzed electronic medical records of 602 patients aged 2-24 months who presented during the bronchiolitis seasons of 2016-2019 with signs and symptoms consistent with bronchiolitis. We excluded medically complex patients, return visits, and those with evidence of alternative or additional pathology such as focal lung findings. All CXRs collected at the general ED were read by GR to direct patient management. CXRs were then re-read by a PR as part of our initial QI study. The PR was blinded to the clinical history and GR read.
Results: Altogether 263 children met inclusion criteria, and 103 underwent CXR (39.1%, 95% CI). GR read 20 pneumonia, 30 viral, 40 normal, and 12 equivocal or “questionable pneumonia” across all seasons. PR read 0 as pneumonia, 91 viral, 3 normal, and 8 equivocal. The interpretation rate of equivocal or pneumonia between GR and PR was 11.7% (95% CI 6.8%, 19.3%) and 19.4% (95% CI 12%, 27%) versus 8.7% (95% CI 4.6%, 15.8%), respectively, with a difference of 10.7% (95% CI 1.1%, 20.3%), P=0.03 (See Figure 1). No pneumonia was read by PR. GR read of pneumonia or equivocal was associated with antibiotics at discharge, with (16/20) 80% (95% CI 58.4%, 91.9%) and (6/12) 50% (95% CI 25.4%, 74.6%) receiving antibiotics at discharge, respectively. Two patients with GR read of pneumonia received antibiotics for pneumonia and acute otitis media. Conclusion(s): Our study reveals discordant CXR interpretations between GR and PR in the setting of bronchiolitis. Given exclusion of patients with focal lung findings, as many as 22 patients may have received unnecessary antibiotic treatment based solely on chest radiograph interpretation by GR. The results of this study highlight the impact of CXR misinterpretation on unnecessary antibiotic exposure, and ultimately reinforce the recommendation against CXR in the workup of bronchiolitis.
GR versus PR CXR interpretations and frequency of antibiotics ordered at discharge. PNA: Pneumonia; AOM: Acute otitis media
Authors/Institutions: Faith Myers, University of Minnesota Masonic Children's Hospital, Minneapolis, Minnesota, United States; Sarah Swenson, University of Minnesota Masonic Children's Hospital, Minneapolis, Minnesota, United States; Ashley Green, University of Minnesota Masonic Children's Hospital, Minneapolis, Minnesota, United States; Dorothy L. Curran, University of Minnesota Masonic Children's Hospital, Minneapolis, Minnesota, United States; Joseph Alfano, Fairview Lakes Medical Center, Wyoming, Minnesota, United States; Nicholas Jubert, Fairview Lakes Medical Center, Wyoming, Minnesota, United States; Kelly R. Dietz, University of Minnesota Masonic Children's Hospital, Minneapolis, Minnesota, United States; Jeffrey P. Louie, University of Minnesota Masonic Children's Hospital, Minneapolis, Minnesota, United States