52. Development and Implementation of a Short Duration of Antibiotic Therapy Algorithm for Uncomplicated Gram-Negative Bacteremia

BACKGROUND: Recent literature suggests no difference in clinical outcomes between short (7 days) and prolonged course (14 days) antibiotic therapy for the treatment of uncomplicated Gram-negative bacteremia (GNB). METHODS: The objectives of the study were to develop and implement a treatment algorithm that identifies patients who are eligible for 7-day therapy for uncomplicated GNB and evaluate its impact on patient outcomes at The Johns Hopkins Hospital (JHH) in Baltimore. The algorithm was developed and implemented at JHH on 11/11/2019. From 11/11/2019 to 3/31/2020, the Infectious Diseases (ID) Pharmacy Resident and ID pharmacists reviewed cases of GNB on weekdays and contacted teams to provide algorithm-compliant treatment recommendations. To quantify the impact of the intervention on clinical outcomes, data from the same time period during the previous year (baseline) were collected and compared to those collected during the intervention. The primary outcome was duration of antibiotic therapy for GNB. Secondary outcomes included: duration of intravenous (IV) antibiotics, length of hospital stay (LOS), and recurrent bacteremia. RESULTS: A total of 345 patients with GNB were identified (142 baseline; 203 intervention) of which 59 and 55 patients met criteria for 7-day therapy, respectively. The Pitt bacteremia score (median 1), bacteremia source [urinary (43%), abdominal (23%)], and organisms [E. coli (48%) and Klebsiella spp. (33%)] were similar between the periods. More patients in the intervention period were treated for ≤8 days (60.0% vs. 37.3%; p=0.015), and the median duration of therapy was 2 days shorter (8 vs. 10 days; p=0.04). Median duration of IV antibiotic therapy (4 vs. 7 days; p=0.004) and median LOS (4 vs. 7 days; p=0.029) were also shorter in the intervention period. There were no differences in the rate of 30-day recurrent bacteremia between the periods (3.4% baseline vs. 1.8% intervention; p=0.60). CONCLUSION: Our pharmacist-led intervention successfully shortened the duration of therapy, increased conversion from IV to PO therapy, and reduced LOS, without negatively impacting the number of patients with recurrent GNB. DISCLOSURES: All Authors: No reported disclosures