2362. Back to the Future: The Impact of Multi-Step Algorithm C. difficile Testing at a Large Tertiary Medical Center

Abstract Background Antibiotic stewardship and infection control programs rely on C. difficile infection (CDI) test results to measure CDI incidence in the hospital setting. C. difficile carriage is common and distinguishing infection from colonization is difficult with the highly sensitive nucleic acid amplification testing (NAAT) commonly used. Current guidelines recommend a multi-step algorithm for testing. The impact on patient outcomes and CDI metrics are largely unknown. Methods This was a pre-post study at the University of Maryland Medical Center, evaluating the impact of a CDI testing strategy (introduced October 2018) that simultaneously reported NAAT and confirmatory enzyme immunoassay (EIA) when used with existing best practice alerts for appropriate testing. Pre-intervention (November 2017–September 2018) and post intervention (October 2018–March 2019) periods were compared for mean CDI incidence (CDI per 10,000 admissions) defined by: (1) positive NAAT, (2) reported CDI (last positive test), and (3) treated CDI (receiving oral vancomycin). Both community and hospital-onset cases were included. The NAAT CDI incidence was used as the pre-intervention comparison for all 3 measures. In addition, oral vancomycin days of therapy (DOT) per 1,000 patient-days (PD) was compared. Pre–post comparisons of mean CDI incidence and mean DOT rates were done using Student t-test. Results There were 3,237 samples tested (2,269 pre and 968 post-intervention) with 376 NAAT positive (262 pre and 114 post-intervention). Of the 99 tests with reflex EIA, there were 74 discordant tests (NAAT +/EIA -) with 35 (47%) treated for CDI. Mean NAAT CDI incidence pre-intervention was 54 per 10,000 admissions. Post-intervention mean CDI incidence decreased as follows: 45 NAAT CDI per 10,000 admissions (P = 0.13), 15 reported CDI per 1000 admissions (P < 0.0001), and 28 treated CDI per 10,000 admissions (P = 0.0007). Oral vancomycin DOT per 1,000 PD decreased from 16 to 9 (P = 0.0002). Conclusion C. difficile NAAT testing with confirmatory EIA, in combination with best practice alert, decreased reported and treated cases of CDI, which may distinguish infection vs. colonization and avoid unnecessary treatment, beyond that achieved with alerts that improve appropriate patient selection for testing. Disclosures All authors: No reported disclosures.