Table 7.

Excess relative risk and relative risk estimates of lung cancer mortality from the interaction models, joint cohort < 100 WLM

Parameter	Deaths	Model A		Model B		Model C		Model D
		Parameter estimatea	95% CI	Parameter estimatea	95% CI	Parameter estimatea	95% CI	Parameter estimatea	95% CI
Background
(β) ERR/WLM	408	0.022	0.013–0.034	0.035	0.018–0.060	0.047	0.020–0.096	0.056	0.020–0.133

				Relative riskb	95% CI	Relative riskb	95% CI	Relative riskb	95% CI
Effect modifiers
Time since exposure (WLM)
5–14 years				1.00		1.00		1.00
15–24 years				0.96	0.50–1.90	0.96	0.50–1.86	0.89	0.45–1.74
25 + years				0.33	0.12–0.70	0.36	0.11–0.76	0.38	0.14–0.78
Attained age (years)
< 55	119					1.00		1.00
55–64	167					0.82	0.30–2.17	0.78	0.29–2.10
65–74	99					0.37	0.02–1.35	0.37	0.04–1.26
75 +	23					0.26	0.00–3.24	0.02	N/A
Exposure rate (WL)
< 0.5	212							1.00
0.5–4.9	182							0.83	0.48–1.66
5.0 +	14							2.00	0.44–5.86

Maximum likelihood estimates and 95% confidence interval

WL working levels, WLM working-level months, ERR/WLM excess relative risk per working-level months, CI confidence interval, N/A not available due to convergence problem in the EPICURE package

^aERR/WLM for cumulative radon exposure, 5-year lagged

^bRelative risks for time since exposure, attained age, and exposure rate variables. Stratification was by sub-cohort (3 categories), and age at risk (12 categories). Parameters are estimated on the basis of the model below. Here the bracketed area represents the cumulative exposures and parameter estimates obtained in different time windows (5–14, 15–24, and 25 + years previously). Subscript a denotes categories of attained age and the subscript z denotes categories of radon concentration in WL

$\text{Rat}{\text{e}}_w=\left[1.0+\mathit{\beta }\left(w_{5-14}+{\mathit{\theta }}_{15-24}{w}_{15-24}+{\mathit{\theta }}_{25+}{w}_{25+}\right)\times {\mathit{\gamma }}_a{z}_z\right]$