Imported Disease Pressure and Bulk Antibiotic Use: Novel Predictors of Facility-Level C. Difficile Infection (CDI) Incidence in the United States Veterans Health Administration

Abstract

Background. Although individual-level risk factors for CDI have been extensively studied, facility-level factors have not. Our objective was to study the determinants of CDI incidence across acute care (AC) and long-term care (LTC) facilities, with a specific interest in the role of imported disease pressure and facility-level antibiotic use. Methods. We conducted a retrospective cohort study of CDI from 2006 to 2012 across Veterans Affairs local healthcare systems where both AC and LTC patient censuses were above an average of 10 patients per day. Our outcome was AC- or LTC-onset C. difficile laboratory-identified event, defined as a case with onset ≥3 days after admission occurring at least 8 weeks from a previous positive test. The facility-level exposures we studied included: imported disease pressure (prevalence of other-facility-onset CDI cases per 10,000 patient-days), antibiotic use (days of therapy per 1000 resident-days), mean Charlson comorbidity count, and mean resident age. Results. 86 LTC and 86 AC facilities met our inclusion criteria. The incidence of C. difficile infection in LTC was 3.6 per 10,000 patient-days in LTC (n = 5968 cases) and was 18.6 per 10,000 patient-days (n = 14,765 cases) in AC. Facility-level antibiotic use was much lower in LTC (median = 137 DOT per 1000 patient-days, range: 61-370) compared to AC (median = 470, range: 378-604, p < 0.001). Conversely, imported disease pressure much higher in LTC (median = 46 per 10,000 patient-days, range: 3-341) compared to AC (median = 11, range: 1-56, p < 0.001). In weighted linear regression and Poisson random-effects models, the most important predictor of CDI incidence was antibiotic use (R2 = 0.52, p < 0.001). Time-series analyses revealed that increases in imported disease pressure preceded increases in CDI rates in LTC for a period of up to 8 weeks, but imported disease pressure was not a significant predictor in AC. Multi-level analyses revealed that importation and facility-level antibiotic use acted independently of individual risk factors, including direct antibiotic exposure. Conclusion. This is the first study comprehensively comparing imported disease pressure and bulk antibiotic use across both LTC and AC facilities. This research points to the need to: (1) monitor and reduce facility-level antibiotic use in both AC and LTC, and (2) bolster infection control programs in LTC facilities where importation is high. Disclosures. All authors: No reported disclosures.