970. Antibiotic Use Variability Among US Nursing Homes—2016

Abstract Background Antibiotics are frequently prescribed in nursing homes (NH). National data describing facility-level antibiotic use (AU) in NH are lacking. The objectives of this analysis were to use NH electronic health records (EHR) to describe AU in NH and variability in AU rates across NH. Methods We analyzed antibiotic orders for 309,884 residents in 1,664 US NHs using one EHR company in 2016. We calculated AU rates as antibiotic days-of-therapy (DOT) per 1,000 resident-days and compared by the type of stay (short-stay (SS) ≤ 100 days vs. long-stay (LS) >100 days). We also examined prescribing indications and the duration of nursing home-initiated antibiotic orders. We assessed facility-level correlates of AU using resident health and NH facility characteristics publically available through NH Compare and LTCfocus using a univariate linear regression. Results In 2016, 57% of NH residents received at least one systemic antibiotic; overall rate of AU was 90 DOT/1,000 resident-days. The median facility-level AU rate was 64 DOT/1,000 resident-days (IQR 36–104). The median proportion of SS residents at a facility was 74% (IQR 60–84%). The SS and LS AU rates were 241 DOT/1,000 resident-days (IQR 173–342) and 24 DOT/1,000 resident-days (IQR 14–37), respectively. Overall, the three most common antibiotic classes prescribed were fluoroquinolones (18%), cephalosporins (18%), and extended-spectrum β-lactams (10%). Antibiotics were most frequently prescribed for urinary tract infections, and the mean duration of an antibiotic order was 9 days (range 1–365). Higher facility AU rate correlated positively with the following facility characteristics; proportion of SS residents, urban location, proportion of residents with mild cognitive impairment and lower activities of daily living scores, presence of ventilator beds, proportion of LS residents with urinary catheters or pressure ulcers, facility case-mix index, and not-for-profit ownership and multiorganization facilities. Conclusion Significant variability in NH AU rates exist, and SS residents have higher AU rates. Identifying NH with high rates of AU after adjusting for facility-level predictors of AU may identify opportunities for targeting efforts to improve prescribing practices. Disclosures All Authors: No reported Disclosures.