Population-based surveillance of Clostridium difficile infection in Manitoba, Canada, by using interim surveillance definitions.

OBJECTIVE: To apply interim surveillance definitions of Clostridium difficile infection (CDI) cases to 1 year of data from the provincewide surveillance system of Manitoba, Canada, to determine the epidemiology of CDI incident cases in a population. METHODS: CDI cases were categorized with interim surveillance definitions developed by an ad hoc C. difficile surveillance working group. Incident cases recorded in the provincial CDI database between July 2005 and June 2006 were linked to the provincial hospitalization and nursing home databases and analyzed. RESULTS: One thousand six incident cases were identified over 1 year. Five hundred fifteen (51%) cases were associated with and began in a healthcare facility (HCF), whereas 275 (27%) were associated with and began in the community. An additional 131 (13%) cases were HCF associated but began in the community, while 85 (8%) were of indeterminate origin. Cases of HCF-associated CDI occurred in patients who were older than did cases of community-associated CDI (P < .0001). The provincial rate of community-onset cases was 23.4 per 100,000 person-years, and rates varied among geographic areas. HCF-associated CDI rates among the 10 largest hospitals varied from 0.5 to 8.4 per 10,000 patient-days. The time to CDI onset after hospital admission indicated that 25% of nosocomial cases began by the 8th day, and 50% began by the 17th day. CONCLUSIONS: Although the majority of CDI cases were associated with exposure to a HCF, 40% of incident CDI began in the community. Populations with HCF- and community-associated CDI demonstrated significantly different age distributions. The wide variation of rates among HCFs requires explanation. The high percentage of incident cases in the community warrants increased study.