Changes of neoplasm concentration with geographical co-ordinates

We investigated the regional differences in diversity--from another respect concentration--of causes of cancer mortality in England and Wales. Statistical analyses of published mortality data were done, using the jackknifed estimate of the Shannon index, for each region, sex and age group and year of death. In males diversity of cancer death causes is secularly increasing whilst it is decreasing in females. Latitude was negatively associated with diversity in male 45+ year age groups and longitude negatively associated with diversity in male 65+ age groups. Although, there were some significant associations in the female groups, there was no general trend across age groups as found in males. These trends remained after accounting for regional variation in past smoking behaviour. We suggest that the observed patterns may be related to prior occupational exposures and non-identified environmental and socioeconomical factors. It is concluded that techniques drawn from population ecology have a potential value in epidemiological studies of human disease. Ecological methods by themselves are likely to be of value in hypothesis generation rather than hypothesis testing.