OBJECTIVES. To compare Clostridium difficile infection (CDI) rates determined with use of a traditional definition (ie, with healthcare-onset CDI defined as diagnosis of CDI more than 48 hours after hospital admission) with rates determined with use of expanded definitions, including both healthcare-onset CDI and community-onset CDI, diagnosed within 48 hours after hospital admission in patients who were hospitalized in the previous 30 or 60 days, and to determine whether differences exist between patients with CDI onset in the community and those with CDI onset in a healthcare setting. DESIGN. Prospective cohort SETTING. Tertiary acute care facility. PATIENTS. General medicine patients who received a diagnosis of CDI during the period January 1, 2004, through December 31, 2005. METHODS. CDI was classified as healthcare-onset CDI, healthcare facility-associated CDI after hospitalization within the previous 30 days, and/or healthcare facility-associated CDI after hospitalization within the previous 60 days. Patient demographic characteristics and medication exposures were obtained. The CDI incidence with use of each definition, CDI rate variability, patient demographic characteristics, and medication exposures were compared. RESULTS. The healthcare-onset CDI rate (1.6 cases per 1,000 patient-days) was significantly lower than the 30-day healthcare facility-associated CDI rate (2.4 cases per 1,000 patient-days; P<.01) and the 60-day healthcare facility-associated CDI rate (2.6 cases per 1,000 patient-days; P< .01). There was good correlation between the healthcare-onset CDI rate and both the 30-day (correlation, 0.69; P< .01) and 60-day (correlation, 0.70; P< .01) healthcare facility-associated CDI rates. There were no months in which the CDI rate was more than 3 standard deviations from the mean. Compared with patients with healthcare-onset CDI, patients with community-onset CDI were less likely to have received a fourth-generation cephalosporin (P = .02) or intravenous vancomycin (P = .01) during hospitalization. CONCLUSIONS. Compared with the traditional definition, expanded definitions identify more patients with CDI. There is good correlation between traditional and expanded CDI definitions; therefore, it is unclear whether expanded surveillance is necessary to identify an abnormal change in CDI rates. Cases that met the expanded definitions were less likely to have occurred in patients with fourth-generation cephalosporin and vancomycin exposure.