Objective: To use aggregated data on smoking habits and lung cancer incidence in occupations assumed to carry no lung cancer risk to control for confounding in other occupational groups.
Methods: Lung cancer incidence was observed from 1971 to 1991 for 53 occupational groups and a group of economically inactive men in a national cohort study. Data on occupation and smoking habits were collected from national surveys during 1965-1980. The relationship between smoking habits and lung cancer incidence was estimated on aggregated level using data from 12 occupational groups that were initially assumed not to be exposed to occupational lung carcinogens. The estimated relationship was used to control confounding from smoking in the other groups. The results were presented as smoking-adjusted incidence ratios.
Results: A significant excess risk was found for 26 groups. It was estimated that about 20% of all lung cancer among men could be related to occupation after adjusting for the effect of smoking.
Conclusions: Our method provided a clearer picture of the occupational risk and could be useful in other situations where individual information on smoking habits is lacking.