185. Implementation of a Clinical Decision Support System (CDSS) for Ordering Antimicrobials in a Veterans Affairs Medical Center (VAMC): Results From a Pilot Quality Improvement (QI) Project

BACKGROUND: When used as antimicrobial stewardship (AS) tool, CDSS built into electronic health record systems (EHRs) have been associated with more appropriate antimicrobial (AM) use. This QI project aimed to determine the perception toward, acceptance and utilization of an AM-CDSS by internal medicine providers at the Houston VAMC. METHODS: From October 2017 to March 2018 (pilot period), ID pharmacists trained members of two to four general medicine inpatient teams/month in the use of the AM-CDSS and provided surveys to be completed during the first (pre) and last week (post) of a 1-month rotation. Surveys focused on the provider’s prescribing patterns, self-perception of infectious diseases knowledge, as well as provider’s awareness, perception of utility, ease of use of the AM-CDSS, and its impact on prescribing patterns. A retrospective chart review was conducted to compare the appropriateness, route and duration of AMs ordered and not ordered through the AM-CDSS. For this comparison, patients were randomly selected from patients with same discharge ID diagnosis matched 1:1 each month. RESULTS: Through the pilot period, the AM-CDSS had a continual rate of increased usage of 4.4 AM orders/month. A total of 113 surveys were collected (63 pre-AM-CDSS and 50 post-AM-CDSS). Eighty percent of respondents reported having used the AM-CDSS, and 76% reported wanting to continue to use the AM-CDSS. The most common reasons cited for using it were to confirm their AM approach and to practice up-to-date evidence-based medicine. Thirty percent stated that it was hard to locate within the EHR. A total of 120 AM orders (AM-CDSS = 60 and non-AM-CDSS = 60) were reviewed for appropriateness. The most common indications for the orders were CAP (34.2%) and UTI (33.3%). AM-CDSS orders were more likely to be appropriate (83.3 vs. 50%, P = 0.0002) and to include oral AMs (58 vs. 13.6%, P < 0.0001) than non-AM-CDSS orders. There was a not significant shorter duration of therapy in the AM-CDSS group (8.7 vs. 9.4d, P = 0.46). CONCLUSION: Most providers perceived the AM-CDSS as useful and easy to use. The use of the AM-CDSS was associated with more appropriate AM use and more frequent selection of appropriate oral AMs. When feasible, AS programs with trainees should consider including easily accessible AM-CDSS in their EHR. DISCLOSURES: All authors: No reported disclosures.