91. Development of an Electronic Algorithm to Identify Inappropriate Antibiotic Prescribing for Pediatric Otitis Media

Abstract

BackgroundAntibiotic stewardship (AS) interventions have primarily focused on acute care settings. The majority of antibiotic use, however, occurs in outpatients. The electronic health record (EHR) might provide an effective and efficient tool for outpatient AS. We aimed to develop and validate an electronic algorithm to identify inappropriate antibiotic use for pediatric outpatients with acute otitis media (AOM).MethodsWithin the Children’s Hospital of Philadelphia (CHOP) Care Network, we used ICD-10 diagnostic codes to identify patient encounters for AOM at any CHOP practice between 3/15/17 – 3/14/18. Exclusion criteria included underlying immunocompromising condition, comorbidities, and concurrent infections that might influence antibiotic use. We randomly selected 450 eligible subjects (150 each from academic practices, non-academic practices, and urgent care). Inappropriate antibiotic use based on CHOP and professional society guidelines were assessed via chart review and served as the basis for assessment of the electronic algorithm which was constructed using only data in the electronic health record (EHR). Criteria for appropriateness focused on the decision to prescribe, the choice of antibiotic, and duration of therapy.ResultsOf 450 subjects, median age was 2, 46% were female, and 88% were evaluated by a physician (vs. advanced practice provider). On chart review, the prescribing decision was correct in 438 (97%), of which 25 appropriately received no antibiotics. Of the 413 subjects who were appropriately prescribed an antibiotic, the choice of antibiotic was appropriate in 37 (9%). Finally, of the 413 patients who were appropriately treated, 412 (99.7%) received the correct duration. Test characteristics of the EHR algorithm (compared to chart review) are noted in the Table.ConclusionFor children with AOM, an electronic algorithm for identification of inappropriate antibiotic prescribing is highly accurate. This algorithm can also highlight for which elements of prescribing the impact of an intervention might be greatest (i.e., choice of agent). Future work should validate this approach in other health systems and geographic regions and evaluate the impact of an audit and feedback intervention based on this tool.Table. Test Characteristics of Electronic Algorithm for Inappropriate Prescribing, Agent, and Duration Disclosures All Authors: No reported disclosures