The National Haemovigilance Office has collected and analysed reports on errors associated with transfusion since 2000. A 3-year pilot research project in near-miss event reporting commenced in November 2002. Near-miss reports from 10 hospital sites were analysed between May 2003 and May 2005. The Medical Event Reporting System for Transfusion Medicine was used to collect and analyse the data. Root cause analysis was used to identify causes of error. A total of 759 near-miss events were reported. Near misses are occurring 18 times more frequently than adverse events causing harm. Sample collection was found to be the highest risk step in the work process and was the first site of error in 468 (62%) events. Of these, 13 (3%) involved samples taken from the wrong patient. Medical staff were frequently involved in error. The general wards and emergency department were identified as high-risk clinical areas, in addition, 78 (10%) events occurred within the transfusion laboratory. Three specific human and two system failures were shown to have been associated with the errors identified in this study. This study confirms that near-miss events occur far more frequently than adverse events causing harm. Collecting near-miss data is an effective means of highlighting human and system failures associated with transfusion that may otherwise go unnoticed. These data can be used to identify areas where resources need to be targeted in order to prevent future harm to patients, improving the overall safety of transfusion.