Intra-operative awareness in paediatric patients has been little studied for many years because of the difficulties in relying on children's testimony. Earlier questionnaires used to detect this complication were not adapted to children's language and memory capacities. By using a qualitative method, a semi-structured in-depth interview adapted to their cognitive abilities, we have now conducted a prospective evaluation of the incidence and risk factors for intra-operative awareness in children undergoing general anaesthesia. Data were obtained from interviews with 410 children (aged 6-16 years) which were conducted within 36 h of general anaesthesia for elective or emergency surgery. One month after surgery, 293 of these patients were interviewed again. Three independent adjudicators classified each potential case of awareness. We considered awareness to include both the 'confirmed awareness' and the 'possible awareness' cases. The accuracy of the children's recall was calculated. The relationship between their awareness and the anaesthesia management was examined. There were five cases of confirmed awareness, and six cases of possible awareness. The incidence of confirmed awareness was 1.2%, but when the possible cases were also considered, the overall incidence of this complication was as high as 2.7% (95% confidence interval, 1.4-5.0%). The only predictive factor identified was the multiple manoeuvres with which the airways were secured (odds ratio, 8.4; 95% confidence interval, 2.4-29.07%). The present study confirms the existence of intra-operative awareness in the paediatric population. The application of a semi-structured in-depth interview adapted to the cognitive capacities of the children appears to enhance the detection of awareness in this population.