Electronic medical records can identify prescribing patterns for antibiotic allergies, study results show

Following analysis of 342 children's electronic medical records, researchers identified 67 children with a possible or probable antibiotic allergy.
Young child with cannula in hospital bed

Prescribing patterns from electronic medical records (EMRs) could identify previously undiagnosed antibiotic allergies in children, researchers have concluded.

An analysis published in the British Journal of Clinical Pharmacology on 20 May 2026 searched EMRs from the Royal Children’s Hospital in Melbourne, Australia, for cases between May 2016 and May 2023 where children had been prescribed adrenaline and/or an antihistamine within 48 hours of receiving a final antibiotic dose.

For their analysis, the study authors assumed that adrenaline and/or antihistamines given shortly after the final antibiotic dose were done so owing to antibiotic allergy. Patients were excluded from the analysis if they presented with eczema, cellulitis or a rash with no documentation of antibiotic-related cause.

Possible allergic reactions were classified as ‘immediate’ (occurring less than one hour after antibiotic administration) or ‘non-immediate’ (occurring after more than an hour). Allergy symptoms were classified as ‘severe’ or ‘non-severe’ reactions.

Of 342 children’s records included in the original analysis, 67 children (20%) were identified as having a possible or probable antibiotic allergy. Ceftriaxone was the most commonly implicated antibiotic, identified in 24 (30%) of cases. Flucloxacillin and amoxicillin were the second and third most common, in 13 (16%) and 12 (15%) cases, respectively.

Of these 67 children, 38 (57%) had already been documented as having an allergy. Just under half of these (43%; n=29) were newly identified as having a possible or probable allergy, “despite the hospital already employing two established processes for identifying suspected allergies: clinician reporting and ICD-coding reports”, the authors said.

Researchers completed an adverse drug reaction report for each child with a newly-identified reaction, all of whom were also given an allergy label on their EMR.

The vast majority of identified allergic reactions (82%; n=55) were identified as ‘non-immediate’ and ‘non-severe’.

According to the study authors, this is the first time an EMR prescription-pattern algorithm has been used to detect suspected antibiotic allergy in children. It is now intended that the algorithm will be run once every three months at the hospital, “to ensure timely identification and referral of patients”.

“As an exploratory study, this approach is intended as a screen tool for identifying children for further evaluation, rather than to diagnose antibiotic allergies or to replace already existing diagnostic pathways.

“This approach could be implemented in hospitals’ antimicrobial stewardship programmes to complement existing methods for identifying those with potential antibiotic allergies who can then have their allergy formally evaluated,” the authors concluded.

Amira Guirguis, chief scientist at the Royal College of Pharmacy, commented: “Accurately identifying and recording suspected antibiotic allergies in children is important for both patient safety and antimicrobial stewardship.

“Some symptoms associated with suspected antibiotic allergies, such as rashes, can overlap with common childhood illnesses and viral infections, so appropriate clinical assessment and follow-up are essential before confirming or excluding an allergy label.

“Inaccurate or incomplete allergy records can affect future prescribing decisions and may lead to unnecessary avoidance of first-line antibiotics.

“Pharmacists play an important role in reviewing medicines histories, supporting safe prescribing, identifying potential discrepancies in allergy records and helping ensure that information is kept up to date across care settings.”

Last updated
Citation
The Pharmaceutical Journal, PJ May 2026, Vol 319, No 8009;319(8009)::DOI:10.1211/PJ.2026.1.413632

    Please leave a comment 

    You may also be interested in