1837. Comparison of Antibiotic Use in Post-Acute and Long-Term Care Facilities Based on Proportion of Short Stay Residents Using a Long-Term Care Pharmacy Database

Abstract

BackgroundCMS requires participating long-term care facilities (LTCF) to have an antibiotic stewardship program (ASP). Common barriers encountered by LTCF include lack of antibiotic use (AU) data and inability to benchmark use. We initiated a project that utilized a long-term care pharmacy (LTCPh) database to obtain and compare AU data across enrolled LTCF.MethodsWe partnered with a regional LTCPh that dispenses and reviews medications for 40 LTCF, of which 32 agreed to participate. Prescriptions filled by the pharmacy were used to calculate antibiotic (AB) starts and days of therapy (DOT). Start and end dates were used to calculate DOT, if available. For those without an end date (<10%), duration was obtained by manual review of administration records. Bed-size and proportion of short-stay (Medicare-A) beds were estimated for each LTCF based on a cross-sectional evaluation of billing records at the LTCPh. Baseline resident-days (RD) during 2017 were obtained from each LTCF. The influence of short-stay residents on AB start rates and DOT was evaluated by grouping LTCF in three cohorts based on estimated proportion of short-stay residents.ResultsData from 29 (90.6%) LTCF were included in the final analysis; 3 were excluded due to lack of RD data. Median bed-size was 57 (range 17–253). Overall, 13.9% of LTCF residents were in the short-stay category. Fifteen LTCF were estimated to have 5% to 20% of RD attributable to short-stay residents, six had <5% while eight had >20%. Antibiotic starts/1000 RD varied from 3.84 to 19.38 and DOT/1000 RD from 34.86 to 252.09, and showed strong correlation (Figure 1). The proportion of short-stay beds correlates better with AB starts/1,000 RD than DOT/1,000 RD (Figure 2). LTCF cohort with >20% short-stay residents had higher mean AB starts/1000 RD compared with LTCF with 5%-20% and <5% short-stay residents (13.08, 9.78, 7.45, respectively; P < 0.05 by one-way ANOVA). However, a similar trend was not noted for DOT/1000 RD (179.30, 128.29, 128.12, respectively; P = 0.12). ConclusionLTCPh can play an important role in supporting ASP in LTCF by providing AU data for benchmarking. Antibiotic use in LTCF is highly variable and may be influenced by the proportion of beds dedicated to short-stay residents amongst other factors.Disclosures S. Bergman, Merck: Grant Investigator, Grant recipient. T. Vanschooneveld, Merck: Grant Investigator, Grant recipient. M. S. Ashraf, Merck & Co. Inc.: Grant Investigator, Research grant.