To compare the complication rates associated with orthopaedic trauma surgery performed by unsupervised and supervised trainees. In our hospital, 6361 orthopaedic trauma operations were performed between 1 January 1998 and 31 December 2002. Data pertinent to the surgeon's supervision and postoperative complications were collected. Elective operations were excluded, as consultants were almost always present. Complication rates ensuing in unsupervised and supervised groups were compared using the Chi squared test. Of 6361 orthopaedic trauma operations performed, 3754 (59%) were by unsupervised trainees of varying experience, whereas 2494 (39%) were by supervised trainees or consultants. In 113 (2%) of the operations, the supervision status was not recorded. The complication rate was significantly higher in the supervised than unsupervised group (5.3 vs 3.3%, Chi squared=15, df=1, p=0.0001). The complication rate was not higher for operations performed by unsupervised trainees than those performed in the presence of a consultant.