Correction to: British Journal of Cancer (2003) 88, 1025–1031. doi: 10.1038/sj.bjc.6600831

The authors have recently notified us of an error in the above paper. It had come to their attention that, due to a coding error, a number of patients who received chemotherapy or radiotherapy were misclassified as having no treatment. The proportions of patients receiving any chemotherapy, radiotherapy and any active treatment were too low as reported. The median percentage and range of patients receiving different treatments within the health authorities are shown below in the revised Table 3.

Table 1 (revised) Median percentage and range of lung cancer patients with each treatment by health authority, 1995–1999

In the revised analyses, there was still evidence of health authority level variation in treatment and survival similar to that shown in Figure 1. The associations of area and health service characteristics are shown in the revised Table 5. There were no material changes in the associations of survival with all area and health service characteristics. In the analysis of determinants of chemotherapy, radiotherapy and any treatment, the associations of treatment with the lung cancer incidence rates were sensitive to correction of the error.

Table 2 (revised) Area and organisational characteristics associated with treatment and survival

The main conclusion of the paper that there is significant geographical variation in the treatment of lung cancer, which may be associated with differential access to oncology services, is still supported.