Identification of antibiotic allergies in children using prescribing patterns in the electronic medical record

BACKGROUND: By age three, 5%-75% of children have a documented antibiotic allergy (AL). Despite this prevalence, recognition remains low because of limited clinician knowledge and time. Our study evaluates a novel approach to identifying AL using electronic medical record (EMR) prescription patterns. METHODS: An EMR report was developed to identify paediatric encounters where adrenaline and/or an antihistamine was administered within 48 h of the last antibiotic dose. The first 100 cases were reviewed to refine exclusion criteria. Exclusions included diagnoses of eczema, or rash; antihistamine/adrenaline given before antibiotic initiation; adrenaline infusion; and antihistamines given after opioids or before blood products or monoclonal antibodies. Case notes from the cohort (2016-2023) were reviewed for suspected AL. RESULTS: The EMR report identified 342 children, with 67 (20%) assessed as having a possible/probable AL. Of these, 38 (57%) had a documented allergy, whereas 29 (43%) were newly identified. Implicated antimicrobials were ceftriaxone (34%), flucloxacillin (19%) and amoxicillin ± clavulanic acid (18%). Immediate-severe reactions accounted for eight cases (12%), including anaphylaxis (9%), angioedema (1%) and respiratory distress (1%). Most reactions (82%) were nonimmediate, nonsevere reactions. One case (1%) was a nonimmediate, severe, serum-sickness-like reaction. Of the 29 newly identified cases, six (21%) were considered too high of a risk to undergo antibiotic challenge, and three (13%) had a positive challenge result. CONCLUSIONS: This is the first study using EMR prescription patterns to identify AL in children. One in five children had a potential AL. This method could complement existing allergy identification practices.