Childhood leukemia around nuclear facilities.
Epidemiologic studies on the health effects of living near nuclear facilities have been rare and, indeed, radiobiological models would not predict any detectable increase in cancer risk to the general public from the very low levels of radioactivity emitted by nuclear installations. Thus the recent evidence suggesting an excess of childhood leukemias in the vicinity of certain nuclear sites in the United Kingdom has generated considerable controversy. To help resolve the uncertainty and enhance interpretability of results, future epidemiologic studies will need to be designed with great care (and within realistic cost limits). This commentary suggests three areas for methodologic consideration: (i) definition and modelling of radiation exposure; (ii) selection of cancer sites and sensitive subgroups; and (iii) use of incidence or mortality data. Specific suggestions for further epidemiologic research are offered as well.