Temporal-spatial distribution of childhood leukemia in New York State. Special reference to case clustering by year of birth

A temporal–spatial analysis was made of childhood leukemia within towns and cities of upper New York State, emphasizing the year of birth as the measure of time. Despite previous observations which have suggested that leukemogenic events occur prenatally or soon after birth, a random birth-year distribution was noted among children developing leukemia before age 5 years. This finding does not provide epidemiologic support for the birth-date liability that might be expected to characterize a viral induction of childhood leukemia. In addition, leukemia under age 15 was found to have a significant tendency to cluster according to year of report to the state cancer registry but evidence is presented to suggest that this finding may be an artifact induced by time-space variation in case registration.