A simulation study of radon and thoron discrimination problem in case-control studies.

In most countries, radon is the dominant contributor among natural radiation sources to the radiation exposure dose of the general population. Numerous case-control studies of residential radon and lung cancer have been conducted using passive radon (Rn-222) detectors. These studies showed that radon may increase lung cancer risk, but most of them did not show a significant risk. Recently it was shown that the readings of passive radon detectors that do not employ thoron (Rn-220) discrimination techniques are affected by thoron. Therefore, we conducted a simulation study to evaluate the possible effect of thoron interference on the estimation of radon-related lung cancer risk. Various assumptions were made based on the number of cases, matching ratio, baseline risk, true radon-related risk, distribution of radon and thoron concentrations, correlation between radon and thoron, and radon detectors. The results suggested that in certain circumstances thoron interference in radon measurements resulted in an approximately 90% downward bias. In addition, the magnitude of the bias increased as the geometric mean and geometric standard error of radon concentration decreased and those of thoron increased. In order to resolve this problem, it is necessary to use passive radon detectors with thoron discrimination techniques in epidemiological studies.