A national comparison between the collocated short- and long-term radon measurements in the United States.

Abstract

Knowing the geographical and temporal variation in radon concentrations is essential for assessing residential exposure to radon, the leading cause of lung cancer in never-smokers in the United States. Tens of millions of short-term radon measurements, which normally last 2 to 4 days, have been conducted during the past decades. However, these massive short-term measurements have not been commonly used in exposure assessment because of the conflicting evidence regarding their correlation with long-term measurements, the gold standard of assessing long-term radon exposure. We aim to evaluate the extent to which a long-term radon measurement can be predicted by a collocated short-term radon measurement under different conditions. We compiled a national dataset of 2245 pairs of collocated short- and long-term measurements, analyzed the predictability of long-term measurements with stratified linear regression and bootstrapping resampling. We found that the extent to which a long-term measurement can be predicted by the collocated short-term measurement was a joint function of two factors: the temporal difference in starting dates between two measurements and the length of the long-term measurement. Short-term measurements, jointly with other factors, could explain up to 79% (0.95 Confidence Interval [CI]: 0.73-0.84) of the variance in seasonal radon concentrations and could explain up to 67% (0.95 CI: 0.52-0.81) of the variance in annual radon concentrations. The large proportions of variance explained suggest that short-term measurement can be used as convenient proxy for seasonal radon concentrations. Accurate annual radon estimation entails averaging multiple short-term measurements in different seasons. Our findings will facilitate the usage of abundant short-term radon measurements, which have been obtained but was previously underutilized in assessing residential radon exposure. Tens of millions of short-term radon measurements have been conducted but underutilized in assessing residential exposure to radon, the greatest cause of lung cancer in non-smokers. We investigate the correlations between collocated short- and long-term measurements in 2245 U.S. buildings and find that short-term measurements can explain ~75% of the variance in subsequent long-term measurements in the same buildings. Our results can facilitate the usage of massive short-term radon measurements that have been conducted to estimate the spatial and longitudinal distribution of radon concentrations, which can be used in epidemiological studies to quantify the health effects of radon.