Cancer registration and studies of incidence by surveys.

A chronological record of the initiation of cancer registers and of special surveys to ascertain cancer incidence, with details of the information asked for in each instance, is presented. Comparisons are made between rates obtained for breast cancer in 12 regions where it is believed that 90% or more of all cancers occurring have been recorded. The average intervals between onset of first symptoms and registration at death for cancers of different sites are considered, and the proportions of cases recorded upon death certificates only; and it is suggested that accuracy of age-specific rates may be improved by using the age at onset of first symptoms as basis rather than age at registration or death, and that "inception rates" so derived are more meaningful.A simple method of standardizing rates for differing age-distribution, and of showing the regressions of incidence upon age, in order to facilitate international comparisons, is described, and the resulting data for 10 regions and 11 sites of cancer are tabulated. The uses to be made of such data in order to search for statistical associations between cancer and social, environmental, dietary and habit factors are discussed.