Adjustment for smoking, alcohol consumption, and socioeconomic status in the California Occupational Mortality Study

This paper presents methods for adjusting for smoking, alcohol, and socioeconomic status in death certificate-based occupational mortality surveillance. The methods were applied in the California Occupational Mortality Study, a statewide study of rates based on 180,000 deaths and census estimates of occupations. For each occupation, levels of smoking, alcohol consumption, and socioeconomic status were estimated using National Health Interview Survey and U.S. Census data, and an empirical Bayes procedure was used to improve the stability of smoking and alcohol estimates for small occupations. Expected death rates for occupations were calculated by modeling rates as a function of age, smoking, alcohol, and socioeconomic status with Poisson regression. The effect of adjustment was usually moderate and in the expected direction, and the adjusted mortality ratios were generally closer to 1.0. Full data on agricultural occupations are presented for illustration.