Beginning in 1988, a question added to the Washington State death certificate asked whether the decedent had smoked during the last 15 years of life. We analyzed death certificate data to evaluate the effectiveness of this question in identifying groups with high smoking rates and occupations with high rates of respiratory disease death among nonsmokers. We obtained statistical death certificate data from the Washington State Department of Health for resident deaths occurring between 1988 and 1991. Analyses included information on age, sex, race/ethnicity, marital status, underlying cause of death, high school graduation, smoking during the last 15 years of life, and occupation. Based on logistic regression analysis, we found that male sex, youth, divorced status, or death from lung cancer, chronic obstructive lung disease, or ischemic heart disease predicted a higher risk of smoking during the last 15 years of life. Hispanic ethnicity, single or widowed status, high school graduation, or death from breast cancer, diabetes, motor vehicle accidents, other accidents, or homicide predicted a lower risk of smoking. In farming occupations, there was an excess number of chronic obstructive lung disease deaths among nonsmokers. Findings from this study suggest that patterns of smoking during the last 15 years of life among decedents can provide useful public health surveillance information. The collection of risk factor information, such as smoking, should be recommended for the U.S. standard death certificate. Questions on smoking should be both simple and answerable by informants who may not have known the decedent for a lifetime. Additional studies on the accuracy of smoking history from the death certificate should be conducted.