Inappropriate antibiotic use is a major clinical problem in hospitals. The initial prescribing choices made by emergency department (ED) providers may be partially responsible for this situation. Unlike hospitalists, ED providers do not have the benefit of learning from their prescribing behaviors because they rarely receive feedback or follow‐up information about the patients after they are sent up to the hospital wards.


In this study, we developed and implemented a hospitalist‐delivered academic detailing intervention to improve ED antibiotic prescribing patterns. X ED midlevei practitioners were recruited to participate in this pre‐/postintervention study at a medium‐sized academic medical center. We retrospectively reviewed 242 antibiotic orders from the preintervention period. Based on the analysis of these orders, we met individually with each ED provider to review prescribing patterns (with most attention on inappropriate orders) as well as current practice guidelines. Following the intervention, the prescribing patterns of the providers were followed for 8 weeks (n = 214 antibiotic orders). The proportion of antibiotics prescribed appropriately before and after the intervention were compared with the x test.


Prior to the intervention 63% of prescriptions were appropriate (95% Cl 57%–69%), and 37% of prescriptions were inappropriate (95% Cl 31 %–43%). After the intervention, 81 % of the prescriptions were appropriate (95% Cl 69%–94%) and 19% were inappropriate (95% Cl 14%–23%), P < 0.01. We also saw a 50% reduction in overall pipercillin/tazobactam orders and a ninefold reduction in inappropriate levofloxacin dosing for urinary tract infections.


A carefully planned hospitalist‐driven audit and feedback intervention can result in behavior change among busy providers in another hospital department, namely, the ED. Our findings suggest that hospitalists may be uniquely positioned to fill yet another quality improvement role in our hospitals.

Author Disclosure:

D. Kiyatkin, none; S. Wright, none; F. Kisuule, none.