181. Limited Effectiveness of an EMR Alert-Based Antibiotic Timeout Procedure in Solid Tumor Cancer Patients

BACKGROUND: Computer-based antibiotic time-outs, in which providers receive automated electronic medical record (EMR) alerts regarding continuation of inpatient antibiotics (Anb), are common stewardship initiatives. We assessed the efficacy of such an intervention in oncology patients (pts), who frequently receive Anb when hospitalized. METHODS: An EMR alert triggered 48 hours after starting vancomycin (vanc), cefepime (cef), piperacillin-tazobactam (pip-tazo), meropenem (mero), and fluoroquinolones (flq) was initiated in a tertiary care hospital in November 2018. To assess the efficacy of the intervention in adults with solid tumor malignancies, demographic, vital sign, laboratory, and treatment data were extracted retrospectively from the EMR. Pts with neutropenic fever, organ transplant, trauma, and cardiopulmonary arrest were excluded. We compared length of therapy [LOT; days of therapy per 1000 patient-days (DOT/1000 pd)] via t-test and incidence rate ratio (IRR) for 3- and 12-month periods preceding and following the intervention. November 2018 was excluded as a washout period. RESULTS: The groups did not differ by age, sex, length of stay, or rate of bacteremia (Table 1). Comparing the 3 months before and after the intervention, neither mean LOT (2.9 ± 0.20 vs 2.6 ± 0.14 DOT/1000 pd, p=0.31) nor rate of Anb use changed (IRR 0.97, p=0.32). However, when considering only the Anb targeted by the intervention, cef usage was 1.4 times higher post- intervention (p=0.002), while use of other Anb was similar (Table 2). Comparing 12 months before to 12 months after the intervention, mean LOT was longer after (0.74 ± 0.018) than before (0.68 ± 0.020 DOT/1000 pd; p=0.03), and Anb use increased (IRR 1.3, p< 0.0001). Specifically, mero (IRR 1.8, p< 0.0001) and cef (1.6, p< 0.0001) were used more frequently after the intervention while none were used less (table 2). Table 1: Study Group Characteristics [Image: see text] Table 2: Antibiotic Use Three Months Before and After, and Twelve Months Before and After, the Intervention [Image: see text] CONCLUSION: Despite wide adoption and efficacy in other populations, an EMR-based Anb time-out did not mitigate the continuation of Anb among inpatients with solid tumors. The intervention may require additional measures, such as an active role for pharmacy, to be effective. However, qualitative studies may also be required to understand why providers are hesitant to limit Anb use in this population. DISCLOSURES: All Authors: No reported disclosures