Purpose: The incidence of adverse drug events (ADEs), preventable ADEs, and potential ADEs was determined using pharmacist surveillance. Drug classes associated with ADEs were also identified., Methods: The study was conducted in a 30-bed hospital ward of a Canadian teaching hospital between April 28, 2003, and May 26, 2003. All patients admitted to the general medicine service were eligible for study enrollment. A pharmacist performed surveillance to identify new or worsening symptoms, critical laboratory values, and medication errors. Surveillance consisted of daily communications with staff, daily chart reviews for all inpatients, and investigation of spontaneous incident reports. Data were collected to describe all identified outcomes. This information was rated independently by two clinicians to determine if the outcome was an ADE, a preventable ADE, or a potential ADE. Descriptive statistics were used to calculate outcome rates, which were reported as events per 100 patient-days., Results: During 543 patient-days of observation, 24 ADEs occurred (4.4 per 100 patient-days), of which 14 were preventable (2.6 per 100 patient-days); 13 potential ADEs also occurred (2.4 per 100 patient-days). Of all ADEs, 3 (13%) were life threatening, 11 (46%) were serious, and 10 (42%) were significant. The 24 ADEs were associated with nine different drug classes. Four drug classes accounted for 17 ADEs (71%): antidiabetic agents, antibiotics, glucocorticoids, and sedatives and hypnotics., Conclusion: Pharmacist surveillance revealed that 4.4 ADEs occurred per 100 patient-days, over half of which were preventable. All preventable and potential ADEs occurred during the ordering and administration stages of medication delivery.