Purpose: The elderly receive many medications which may have adverse effects. Little evidence is available about the epidemiology of potentially inappropriate medications being prescribed to the elderly in Japan as defined by the Beers criteria, or whether or not these medications result in harm when used in this population. Methods: We conducted a prospective cohort study of patients aged ≥65 years who were admitted to three acute care hospitals in Japan. Trained research nurses followed up patients from randomly selected wards and collected data about their medications and all potential adverse drug events (ADEs). Two independent reviewers evaluated all the data. The use of potentially inappropriate medications and their effects on patients were identified using the updated Beers criteria. Results: A total of 2155 elderly patients were eligible; 56.1% received at least one drug listed in the Beers criteria (BL drug). The rates of BL drug prescriptions were 103.8 per 100 admissions and 53.7 per 1000 patient-days, and the incidence rate of ADEs related to BL drugs was 1.7 per 100 BL drug prescriptions. Among patients aged ≥65 years, relatively younger patients (p=0.0002) and those with less complications (p=0.04) were likely to be prescribed BL drugs. Conclusions: Although BL drugs were frequently prescribed to elderly Japanese inpatients, the incidence of related ADEs appeared infrequent. These data suggest that re-evaluation of the appropriateness of the Beers criteria is needed before they are used in Japan and other nations to assess quality or for decision support. © 2011 John Wiley & Sons, Ltd.