BACKGROUND: Removing inaccurate penicillin allergy labels (PALs) can reduce unnecessary exposure to 'watch' and 'reserve' groups of antibiotics and thereby reduce antimicrobial resistance. The most efficient model for a non-allergy-specialist-led penicillin allergy de-labelling (PADL) service has not been established. OBJECTIVE: To determine the costs to the UK National Health Service of a direct oral penicillin challenge (DPC) for low-risk patients with a PAL in three hospitals in England, each with a different non-allergy-specialist delivery model: pharmacist-led, nurse-led, and mixed multidisciplinary. METHODS: Cost analysis of the DPC pathway, including resources related to staff time and antibiotics. The effect of de-labelling on healthcare utilisation over 5 years was modelled using data from the published literature. RESULTS: In total, 2257 patients from the Acute Medical or Infectious Disease Unit (AMU/IDU), Pre-surgical, and Haematology-Oncology departments were screened. Subsequently, 126 underwent DPC, and 122 were de-labelled. Twenty-two of these were de-labelled in time to affect their antibiotic regimen
6 from AMU/IDU and 16 Pre-surgery. The DPC represented 22%-23% of the pathway cost in the pharmacist-led and mixed models, and 15% in the nurse-led model. Across departments and models, the cost per de-labelled patient varied between £577 (95% Credible Interval: 370, 633) for haematology-oncology patients to £2329 (947, 19,504) for AMU/IDU patients, both under the nurse-led model. After 5 years, recouping costs was unlikely for AMU/IDU patients under any model or for all patients combined under the mixed model. CONCLUSIONS: The penicillin allergy de-labelling pathway cost was ≥ 4-fold that of the DPC alone. Costs were up to 3 times higher in an acute compared to an elective setting. No short-term cost savings were identified from proactive or opportunistic penicillin allergy de-labelling in this study.