In the box titled “Vaccination with autologous tumour cells” in this article by G A Chung-Faye and D J Kerr (2 December, pp 1397-9) some results from Vermorken et al (1999) were wrongly reported. The second bullet point should read: “In those receiving vaccination, there was a significant reduction in recurrence (44%, 95% confidence interval 7% to 66%) and an increase [not reduction] in overall survival, although this did not reach significance.” Also, on page 1397, the graph showing relative risk of colorectal cancer with use of non-steroidal anti-inflammatory drugs somehow lost its vertical line at x=1 indicating treatment equivalence.