Disagreement Between Allergist and Electronic Health Record Documentation of Penicillin Allergies

Approximately 10% of US patients are labeled as having a penicillin allergy, yet less than 1% have a true allergy.This discrepancy results in unnecessary avoidance of -lactam antibiotics, leading to worse clinical outcomes, increased rates of Clostridioides difficile infections and multidrug-resistant organisms, greater health care costs, and longer hospital stays. 1 Penicillin allergy labels in the electronic health record (EHR) frequently serve as the basis for characterizing, classifying, and riskstratifying reactions; however, these labels are often incomplete, inaccurate, or erroneous.These allergy entries often persist as the primary, and sometimes only, record of a patient's reaction history carried forward through their medical care.In this issue of JAMA Network Open, Novotny et al 2 evaluate the accuracy of penicillin allergy labels by performing a retrospective study of 230 patients with penicillin allergy labels seen in a university-affiliated allergy clinic and comparing EHR documentation with the allergists' history assessment.The authors found disagreement in approximately half of the cases assessed, including inappropriately labeled high-risk penicillin allergies.In this study, Novotny et al 2 compared EHR documentation of a penicillin allergy with allergist assessments of both the primary index reaction symptom and the index reaction risk level.The former was defined by the most severe symptom, such as a severe cutaneous adverse reaction, anaphylaxis, or angioedema, whereas the latter was assigned as low, moderate, or high.In the EHR, risk level was either automatically assigned based on symptom selection within the allergy EHR module or defaulted to low risk if no flag were applied.Allergist risk level was determined using the PEN-FAST criteria.