973. Inter-facility patient sharing and Clostridium difficile incidence in the Ontario hospital network: a 13-year longitudinal cohort study of 116 hospitals
Session: Oral Abstract Session: Healthcare Epidemiology: The Poop Pager and Other Novel Perspectives on C. difficile in the Healthcare Setting
Friday, October 5, 2018: 10:45 AM
Room: S 156


Inter-facility patient movement plays an important role in the dissemination of antimicrobial resistance and C. difficile infection (CDI) throughout healthcare systems. However, the relative performance of different patient sharing metrics for predicting CDI incidence is not known. We compared 3 different measures of inter-facility patient sharing as they relate to CDI incidence in Ontario facilities.


A retrospective cohort analysis was used to predict incident CDI (ICD-10 = A04.7 identified from Discharge Abstract Database records) across Ontario hospitals (Nhospitals=116) between April 1, 2003 to March 31, 2016. Patients with a stay of < 3 days and those with a history of CDI in the prior 90 days were excluded from the risk set but not from patient sharing metrics. Poisson regression models with facility-level random effects were used to predict facility CDI incidence (per 1,000 admissions) and measure the percent change in facility-level variance (PCV). The 3 metrics of inter-facility patient sharing included: (1) ‘importation’ – the rate of patients with a discharge from another distinct facility in prior 90-days, (2) ‘incidence-weighted importation’ – equal to importation weighted by the incidence of CDI in the previous facility, and (3) ‘case importation’ – importation of patients with a history of CDI.


Over the 13-year period, we observed 58,427 cases of healthcare-associated CDI among 12,750,000 admissions. Facility CDI incidence ranged from 2.9 to 19.6 per 1,000 admissions (6.8-fold range). Patient sharing metrics were strongly related to facility CDI incidence (Figure). In models adjusting for facility risk factors, all 3 measures still explained an important portion of inter-facility variation in CDI incidence: importation (PCV=5%, p=0.01), incidence-weighted importation (PCV=15%, p<0.001), and ‘case importation’ (PCV=48%, p<0.001).


We observed a substantial variation in facility CDI incidence that was explained by linkages between acute care facilities, especially linkage to other facilities with a high incidence of CDI. Facility infection prevention staff should consider incorporating the facility CDI incidence into risk stratification assessments of patient transfers.


Kevin Brown, PhD1, Nick Daneman, MD, MSc2, Kevin Schwartz, MD, MSc3, Bradley Langford, BScPhm ACPR PharmD BCPS4, Jennie Johnstone, MD, PhD5, Kwaku Adomako, MSc3, Jacob Etches, PhD6 and Gary Garber, MD, FACP, FIDSA7, (1)Dalla Lana School of Public Health, University of Toronto, Toronto, ON, Canada, (2)Division of Infectious Diseases & Clinical Epidemiology, University of Toronto, Toronto, ON, Canada, (3)Infection Prevention and Control, Public Health Ontario, Toronto, ON, Canada, (4)St. Joseph's Health Centre, Toronto, ON, Canada, (5)Public Health Ontario, Toronto, ON, Canada, (6)Institute for Clinical Evaluative Sciences, Toronto, ON, Canada, (7)University of Ottawa, Ottawa, ON, Canada


K. Brown, None

N. Daneman, None

K. Schwartz, None

B. Langford, None

J. Johnstone, None

K. Adomako, None

J. Etches, None

G. Garber, None

Findings in the abstracts are embargoed until 12:01 a.m. PDT, Wednesday Oct. 3rd with the exception of research findings presented at the IDWeek press conferences.