29.6% of post-operative films (17%–39%) were judged to have sufficient landmarks visible to enable measurement of vertebral rotation compared to 10% of pre-operative films. Marked increase in systematic bias between consultants with post-operative radiographs to pre-operative films was observed.