Abstract
Significant uncertainty is known to be present in the internal radiation dose assessments for intakes of uranium compounds. The aim of this thesis was to investigate the sources of the uncertainty in the dose assessment and in particular the biokinetic models used to describe the movement of uranium in the body. A variety of numerical and statistical approaches have been used to explore the uncertainty in the dose assessments. In particular Bayesian techniques have been applied to historical uranium in urine measurements to derive best estimates for important biokinetic parameters. Posterior probability distributions for the important lung solubility parameters have been calculated from six historical exposure cases. It was shown that improved estimates of these parameters can be obtained by sharing the information in each of the cases in the Bayesian calculation. Using these estimates, the uncertainty in the radiation doses from occupational intakes of uranium have been quantified. Lastly, a novel Bayesian approach to identifying intakes of uranium from bioassay measurements has been developed and tested. This Bayesian approach provides a number of advantages as it incorporates the uncertainty in the bioassay measurements, intake parameters and the bioassay model parameters in a coherent way.