539. Canaries in a Coal Mine?: Early Identification of Regional Spread of Novel Multidrug-resistant Organisms (MDROs) Using Sentinel Surveillance in Skilled Nursing Facilities Caring For Ventilated Patients (vSNFs)

BACKGROUND: Regional containment of novel or targeted MDROs depends on detecting their presence as soon as possible following their introduction. Prior modeling studies suggest that after importation to a region, novel MDROs appear relatively quickly in certain high-risk post-acute long-term care facilities via patient movement. Sentinel surveillance in such facilities might facilitate early detection of emergent MDROs, thereby enhancing the effectiveness of containment efforts. METHODS: We simulated the introduction and spread of carbapenem-resistant Enterobacteriaceae (CRE) in a region using an adaptation of a previously described susceptible-infectious-susceptible model (Clin Infect Dis. 2019 March 28 doi: 10.1093/cid/ciz248). The model includes the patient sharing network among healthcare facilities in an exemplar US state, using claims data and the Minimum Data Set from the Centers for Medicare & Medicaid Services for 2015. Disease progression, transmission and testing rates were estimated for CRE using data from the literature. Each simulated outbreak was initiated with a single importation to a Dartmouth Atlas of Health Care hospital referral region. The predicted timing of first CRE detection using two different data sources was compared: (1) real-time monitoring of clinical microbiology test results, or (2) results from quarterly point prevalence colonization surveys (PPSs). For each data source, the timing of earliest detection was compared according to availability of data from: (a) all healthcare facilities statewide, (b) only long-term acute care hospitals, (c) only vSNFs, or (d) only the largest acute care hospitals in the state (n = 23). RESULTS: Compared with real-time monitoring of clinical microbiology testing results from all facilities statewide, quarterly PPSs at all facilities detected CRE 446 days (median; range 312–608 days) earlier, while PPSs at only vSNFs (representing 4.4% of inpatient beds statewide) detected CRE 385 days (range 194–553 days) earlier (figures). CONCLUSION: Regular point prevalence surveys in vSNFs may detect new MDROs in a region approximately one year sooner than real-time monitoring of clinical microbiology results, and may be an efficient strategy for early regional detection and subsequent containment. [Image: see text] DISCLOSURES: All authors: No reported disclosures.