ICRP (1987) developed its models for environmental risk on the basis of both a constant-relative-risk model and a constant-absolute-risk model. ICRP assumed that the risk expressed over the years that cancer occurs would be increased if exposure occurred in childhood. It assumed that the risk was 3 times as great for exposure at ages 0–20 than for exposure at ages over 20. There is little justification for that assumption, as later information suggests that those exposed as children might have no different risk than those exposed as adults (Lubin and others 1995; Xuan and others 1993). This is discussed further later in this chapter.
The constant-absolute-risk or constant-relative-risk model is no longer considered appropriate for lung cancer. The best models use modifications of the parameters to account for a risk reduction with time since exposure. Although not biologically correct, risk estimates calculated with a constant-relative-risk model are within a factor of 3 of those calculated with other models.
Values of lifetime risk for the ICRP model are shown in table 5.5.
The fourth National Research Council Committee on Biological Effects of Ionizing Radiations (BEIR IV) prepared a report, Health Risks of Radon and other Internally Deposited Alpha-Emitters (National Research Council 1988). The committee was given the raw data or selected parts of the original data from four mining cohorts: the US (Colorado), Canadian (Ontario and Eldorado), and Malmberget (Swedish) cohorts. Reanalysis was performed with the AMFIT program developed for analysis of the Japanese atomic-bomb survivor data. The program uses Poisson regression to estimate parameters.
With AMFIT the data were analyzed with both internal and external cohorts for a control population. The BEIR IV committee stated that a relative-risk model fit the observed mortality well. The relative-risk model assumes that radon decay product exposure increases the age-specific lung-cancer mortality rate in the population by a constant fraction per WLM of exposure. However, in all cohorts, there was an obvious reduction in lung-cancer relative risk with time after exposure. The relative-risk model was modified to reduce risk with time since exposure. The BEIR IV committee called its modified relative-risk model a time-since-exposure (TSE) model.
Smoking was examined as a confounder. The only study with complete smoking history on the miners was the Colorado study. The effect was tested with a hybrid relative-risk model that incorporated a mixing parameter for smoking. A parameter value of zero fit an additive effect of smoking and 222Rn interaction; a value of 1 fit a multiplicative model best. A maximum log-likelihood test was applied to the data, and it was found that the best parameter fit was between 0 and 1. This indicated that combined risk was more than additive but less than multi-