Uncertainties associated with assessing Ontario uranium miners' exposure to radon daughters

The Ontario uranium miners study is a large (n = 28 546) cohort with low levels of radon exposure relative to other uranium miner cohorts. Multiple methods were used over time to estimate annual occupational exposure to radon daughters including: mine-specific extrapolations by mining engineers, area sampling in limited areas of the mines combined with approximate working time and lastly, consistent exposure sampling in different locations of the mine combined with workers' time cards. Nonetheless, estimating exposures involves assumptions that lead to some uncertainty in occupational exposure characterisation arising from the assessment approach and variability within workplace, over time and by individual. An evaluation of the total uncertainty associated with radon daughter exposure estimation in Ontario miners over time has not been conducted. The objective of this study was to identify the contributing sources and assess the total uncertainty associated with estimating occupational radon daughter exposure among underground Ontario uranium miners over the course of uranium mining. The five sources of radon daughter exposure uncertainty evaluated were: natural variations in radon concentration, estimation of working time, precision of the radon measurement method, unintended errors during sampling, and record keeping and transcription of exposure data. These sources were examined separately for the period 1958 to 1967 and then 1968 onward due to changes in radon daughter concentration measurement practices between these periods. The magnitude of uncertainty associated with each of these sources over time were determined by reviewing historical literature on uranium mining in Ontario as well as through expert advice. Using the root sum square method, the total radon daughter exposure uncertainty was found to be 53 to 67% in the earlier period of uranium mining from 1958 to 1967. This decreased to 31 to 39% for the period 1968 to 1996 with natural variations of radon daughter concentrations in mines accounting for the largest percentage of uncertainty. This assessment provides an initial step in understanding the effect of exposure uncertainty on risk estimates. The impact of this uncertainty on the dose-response relationship between radon exposure and cancer risk will be assessed in future work.