Previous studies point to a lower use of radiotherapy by Australian cancer patients in lower socioeconomic areas and in country regions that are some distance from urban treatment centres. These were cross-sectional studies with the potential for error from changes in place of residence. We used a cohort design to avoid such error. South Australian patients diagnosed in 1990−1994 were followed until the date of censoring of 31 December 1999 using data from the State Cancer Registry. The percentage found to have had megavoltage therapy in the first 12 months following diagnosis varied by leading primary incidence site from 44% for the prostate to 40% for female breast, 38% for lung, 17% for rectum, 3% for colon and 2% for skin (melanoma). Multivariate analysis indicated that determinants of not receiving megavoltage therapy in the first 12 months were older age, female sex, residence in a country region and country of birth. Melanoma data revealed earlier stages for women than men. If this difference by sex applies to other cancers, it might explain the lower exposure of women to radiotherapy. Fewer older patients received radiotherapy, consistent with trends observed in hospital-based cancer-registry data. The influence on this finding of differences in stage and comorbidity requires additional study. While earlier findings of a lower exposure of country residents to radiotherapy were confirmed, the difference was comparatively small in this study. Variations in exposure by socioeconomic status of residential area were not observed.