BACKGROUND: Clostridium difficile-associated diarrhea (CDAD) causes substantial healthcare-associated morbidity. Unlike other common healthcare-associated pathogens, little comparative information is available about CDAD rates in hospitalized patients. OBJECTIVES: To determine CDAD rates per 10,000 patient-days and per 1,000 hospital admissions at 7 geographically diverse tertiary-care centers from 2000 to 2003, and to survey participating centers on methods of CDAD surveillance and case definition. METHODS: Each center provided specific information for the study period, including case numbers, patient-days, and hospital characteristics. Case definitions and laboratory diagnoses of healthcare-associated CDAD were determined by each institution. Within institutions, case definitions remained consistent during the study period. RESULS: Overall, mean annual case rates of CDAD were 12.1 per 10,000 patient-days (range, 3.1 to 25.1) and 7.4 per 1,000 hospital admissions (range, 3.1 to 13.1). No significant increases were observed in CDAD case rates during the 4-year interval, either at individual centers or in the Prevention Epicenter hospitals as a whole. Prevention Epicenter hospitals differed in their CDAD case definitions. Different case definitions used by the hospitals applied to a fixed data set resulted in a 30% difference in rates. No associations were identified between diagnostic test or case definition used and the relative rate of CDAD at a specific medical center. CONCLUSIONS: Rates of CDAD vary widely at tertiarycare centers across the United States. No significant increases in case rates were identified. The varying clinical and laboratory approaches to diagnosis complicated comparisons between hospitals. To facilitate benchmarking and comparisons between institutions, we recommend development of a more standardized case definition.