The population risk of radon-induced lung cancer is assessed by an attributable risk (AR). The AR of lung cancer due to ionising radiation is defined as the proportion of lung-cancer deaths attributable to indoor radon exposure(5)
. This risk indicates the proportion of lung cancer deaths that could be theoretically prevented by reducing indoor radon concentrations to outdoor levels. Given the exposure–response relationship for radon and lung cancer risk, the mortality rates of lung cancers and all cases and the distribution of indoor radon concentrations, the AR can be estimated:
is the probability density of radon concentration and
is the lifetime relative risk of lung cancer for a lifetime exposure to radon at a constant concentration of C
in the presence of competing risks, i.e. the ratio of lifetime lung cancer risk for exposure rate C
to lifetime lung cancer risk for ‘zero’ exposure.
The proportions of Canadian homes with different ranges of radon concentration are based on Health Canada's national residential radon survey in homes roughly uniformly distributed in 121 health regions across Canada(4)
. The survey was conducted over 2 y during the heating seasons of 2009/10 and 2010/11. Long-term radon measurements were performed in all the homes surveyed. The observed radon concentration in Canadian homes follows a log–normal distribution(4)
with a population-weighted GM of 41.9 Bq m–3
and a GSD of 2.8. It was assumed that males and females were distributed equally in the various radon concentration ranges.
The lifetime relative risk,
at a given radon concentration, C
, can be calculated from the exposure–response relationship for radon and lung cancer risk. Based on worldwide epidemiological studies, the BEIR VI committee recommended two models for estimating radon risks: the exposure–age–duration model and the exposure–age–concentration model(6)
. The two models are equally preferred and it is not easy to decide which one to use in practice. The United States Environmental Protection Agency devised a single model(7)
that gives risk values midway between the two BEIR VI preferred models. The Environmental Protection Agency model was used in the calculation of lifetime RRs for Canadians(8)
. The computation of lifetime risks depends on the choice of the background age-specific lung-cancer and overall mortality rates. In the calculation, Canadian age-specific mortality rates averaged over 5 y from 1996 to 2000(9)
were used. In the adjustment of age-specific lung cancer mortality rates to reflect smoking status, Canadian age-specific smoking prevalence data for males and females in 2002 were used(10)
. The average age of commencement of smoking is 18 among Canadians. According to the BEIR VI report, never-smokers are defined as those persons who have not yet smoked 100 cigarettes, and ever-smokers are those who have smoked at least 100 cigarettes in their lifetime. It is accepted that smoking and radon exposure combine in a fashion that is sub-multiplicative on the RR scale. It is assumed that smoking-induced lung cancer has a 10-y latency period and the RR for ever-smokers compared with that for never-smokers is approximately 14 for males and 12 for females. Details of the lifetime RR calculation can be found in a previous publication(8)
. It is further assumed that the lifetime
at a given radon concentration C remains unchanged from the previous assessment. Those lifetime RRs are used in the current re-assessment of radon-induced lung cancer with more accurate radon distribution characteristics obtained from the recent cross-Canada residential radon survey.