Implementation and performance of a nurse administered modified PEN-FAST clinical decision rule in the electronic health record

Objective: To evaluate performance of registered nurse assessments of the PEN-FAST penicillin allergy clinical decision rule compared to antimicrobial stewardship pharmacists. Design: Prospective, blinded, non-interventional, quality assurance study. Setting: This study took place across 4 inpatient hospitals within a large health system in Houston, Texas. Methods: We implemented PEN-FAST rule questions into the electronic health record (EHR) for registered nurses to perform. Patients were randomly selected in a prospective fashion, with nurse documented scores hidden, for re-assessment by antimicrobial stewardship pharmacists to compare risk stratification and scores. Results: Overall agreement of high risk and low risk results was 84.3%. Registered nurse evaluations with the PEN-FAST clinical decision rule for detecting a high-risk patient demonstrated a sensitivity of 67%, specificity of 89.8%, positive predictive value of 67.9%, and negative predictive value of 89.5%. Additionally, 34.4% of patients with a documented penicillin allergy admitted to tolerating amoxicillin or amoxicillin/clavulanate since their last recalled reaction to penicillin. Conclusions: Registered nurse assessment of the PEN-FAST clinical decision rule demonstrated good performance and can effectively be used to screen for low-risk penicillin allergy patients. Incorporation of the PEN-FAST rule into EHR can be scaled into large health systems to help appropriately stratify patients with low- and high-risk penicillin allergies and improve documentation.