Evaluation of several methods for assessing the effects of occupational exposure to radiation

The evaluation of health effects in populations occupationally exposed to low-level ionizing radiation is a matter of considerable current controversy. The analysis of data on such exposures presents a variety of problems resulting from the time dependent nature of the exposure data, certain selective biases found in working populations, and particularly limits imposed by the size of the populations, and the magnitudes of exposures received. In this paper, several methods of analysis are presented and evaluated using data from the Hanford plant for illustration. Questions of interest include whether or not to utilize an external control, and how to handle the highly skewed exposure data most effectively. Expressions for the power of various procedures are used not only to compare methods but also to evaluate the potential for detecting effects in occupationally exposed populations