Soviet-American cooperation on fundamental problems in environmental health.

Records of 2,145 cases of Hodgkin's disease in England and Wales treated by the British National Lymphoma Investigation during 1970-84 were sought in the national and regional cancer registers. One thousand eight hundred and eight-six (88%) were recorded in the national register, either as Hodgkin's disease (86%) or as other or unspecified lymphoma (2%) and 2 (0.1%) were recorded as other cancers. A further 69 (3%) cases were registered by regional cancer registries but had not reached the national register. Adjusting for the distribution of the study cases by region of incidence, we estimate completeness of registration of cases of Hodgkin's disease in the national register at 89.7%, and in the regional registers overall at 92.9%. Completeness did not vary appreciably by age or sex or calendar period. There was however, substantial variation in completeness between regional registries. Estimates were made for all regions except North Western; the lowest estimated completenesses were under 90% in Wessex, and the Thames registry regions, and the greatest were 95% or more Because these results are confined to one malignancy treated by a particular collaborative network of physicians (although a large and widespread one), and because the patients are restricted to those seen in hospitals, caution must be exercised in extrapolation of the findings to cancer registration generally, but other studies and sources of information lead to similar conclusions about completeness of cancer registration nationally and regionally. Cancer registration has been conducted in parts of England and Wales since the 1920s, with notionally complete geographic coverage of the country since 1962. The data are collected by regional cancer registries; there were 74 such registries in 1950, diminishing in number over time, such that now there are 12. Since 1945 the regional registries have sent data to the national registry (now at the Office of Population Censuses and Surveys (OPCS)), where they are collated, and validation and elimination of duplicates is undertaken. Multiple sources of data are used to ascertain cancers, with the particular sources varying to some extent between regions (Swerdlow, 1986). All receive extracts from all deaths certificates for residents of their region which mention cancer. The England and Wales data set of over 5 million cancer registrations is exceptionally large, but its use to examine secular and cohort trends in cancer incidence and to provide incidence data in cohort studies, has been hampered by uncertainty about the degree of completeness of registration. …