The objectives of this study were first to develop and institute a methodology for the study of anaesthetic outcome for parallel use in four teaching hospitals in Canada and second, to compare rates of morbidity and mortality associated with anaesthesia between the four centres. The basic design of the study was occurrence screening with anaesthetists entering data on patient demographics, anaesthetic and surgical factors. Research nurses reviewed anaesthetic records and hospital charts and interviewed patients postoperatively. Data on 37,665 anaesthetics were collected during 1988-89 in the four teaching centres. There were major differences found across the hospitals, particularly with regard to volume, patient case-mix, anaesthetic drugs and monitoring used. The use of parallel training, repeated consultations and use of rounds and inservices contributed to the reliability and validity of the data collection. We conclude that outcome surveillance can be instituted in different hospital Departments of Anaesthesia with sufficient confidence to form the basis of comparison of anaesthetic outcome.