RESUMO
BACKGROUND: Job-exposure matrices (JEMs) are often used for exposure assessment in occupational exposure and epidemiology studies. However, general population JEMs are difficult to find and access for workers in the United States of America. OBJECTIVE: We aimed to use publicly available information to develop a JEM-like exposure assessment method to determine exposure to a wide range of occupational agents in a wide range of occupations for US general population studies. METHODS: We used information from the US Department of Labor's Occupational Information Network database (O*NET) for 19,636 job tasks and 974 civilian occupations. We used automated keyword searches to identify 1,804 job tasks that involved exposure to 50 occupational agents. We had 2 reviewers determine whether each identified job task actually involved exposure to the 50 occupational agents. We had a third reviewer, a certified industrial hygienist, assess any job task and exposure for which the first 2 reviewers disagreed. For each U.S. Census 2010 occupation code, we used this information to derive 3 exposure variables for each occupational agent: ever exposure, number of job tasks of exposure, and frequency of exposure. RESULTS: Our keyword searches identified a median of 10 (interquartile range [IQR]: 43.75) job tasks for each occupational agent, and the maximum was 308. We determined job tasks actually involved exposure to 45 occupational agents, including solvents, air pollution, pesticides, radiation, metals, etc. We derived the 3 exposure variables for these 45 occupational agents for 516 U.S. Census 2010 occupation codes. The median percentage for ever exposure to individual occupational agents was 1.16% (IQR: 1.74%), and the maximum was 11.43%. CONCLUSIONS: Our JEM-like exposure assessment method based on O*NET information can be used to determine exposure to a wide range of occupational agents in a wide range of occupations for the US general population.