OBJECTIVES: To evaluate the effect of patients’ documented penicillin allergy on the likelihood of primary care clinicians’ prescribing penicillin-based antibiotics for treatment of common infections.

METHODS: We extracted all outpatient-visit prescriptions from the electronic prescribing system of 63 primary care clinic facilities in Dongcheng district in Beijing city between 2020 and 2022. We included visits associated with at least one antibiotic prescription and a documented diagnosis of one or more of the five most common acute primary care infections for which antibiotics are not routinely recommended but, when necessary, penicillin is the first-line therapy, including pharyngitis, sinusitis, tonsillitis, unspecified upper respiratory tract infection and bronchitis. We excluded visits with other documented infections. Adults were categorized as only having a documented penicillin allergy or no documented antibiotic allergy; patients with a non-penicillin antibiotic allergy were excluded. We then estimated generalized linear models with logit link function to examine the adjusted associations between prescribing of penicillin-based antibiotics and patients’ documented penicillin allergy, controlling for patient age, gender, and intra-cluster correlation among each clinician.

RESULTS: A total of 62,789 outpatient visits were included. Primary care clinicians were less likely to prescribe penicillin-based antibiotics when treating common infections among patients with documented penicillin allergy when compared with patients without any documented antibiotic allergy (5.2% [156/2,979] vs. 7.9% [4,740/59,810], p<0.001, OR 0.67 [0.56-0.80]). Moreover, patients with penicillin allergies were less likely to be prescribed second-generation cephalosporins (44.1% [1,314/2,979] vs. 50.6% [30,273/59,810], p=0.002, OR 0.76 [0.64-0.90]), but more likely to be prescribed macrolides (24.7% [736/2,979] vs. 13.2% [7,903/59,810], p<0.001, OR 2.18 [1.85-2.57]) and fluoroquinolones (8.4% [251/2,979] vs. 6.2% [3,738/59,810], p<0.001, OR 1.41 [1.18-1.68]).

CONCLUSIONS: Pen-A documentation could significantly affect primary care clinicians’ use of first-line penicillins in China. Interventions that support antibiotic allergy assessment and de-labeling of patients in primary care may improve optimal antibiotic therapy.