Background: There is a potential conflict of interest in providing the best possible outcome for patients undergoing cardiac surgery and good training for junior cardiac surgeons. Methods: We analysed training with reference to volume of work, risk stratification and outcome for consultant and trainee procedures. The Parsonnet system was used for risk stratification. A retrospective audit was performed for a 6-year period. Results: During the study period, 6037 operations were performed, of which 2166 were carried out by trainees. Direct consultant assistance in a trainee operation varied between 17 % and 51 % and increased towards the end of the study period. Of the operations performed by trainees, 88 % were CABG. The median Parsonnet score for consultant operations was 9 compared with 4 for trainees. Actual mortality was below predicted for all surgeons. Morbidity was also lowest for trainees (10 %). Conclusions: With appropriate case selection, trainees in cardiac surgery can achieve good results. As training changes in the UK, trainees should receive increased supervised exposure to a wider range of procedure to compensate for a lower volume of workload. 1 This work was presented at the Symposium on Risk Stratification in Cardiac and Thoracic Surgery in Cologne on October 15 - 16, 1999