A triangulation approach to historical exposure assessment for the carbon black industry

The determination of cumulative exposures for individual workers is necessary for research and practice of occupational health and hygiene. Reconstruction of exposures for a study of respiratory morbidity was needed to study the effects of exposure to carbon black production. Approximately 15,800 exposure estimates were needed. There were 22 plants, a 40-year time span, six job categories, and three types of dust-exposure metrics (respirable, inhalable, and total dust). Three information sources were used: 1) Industrial hygiene air level measurements where available (several industry-wide surveys had been conducted). 2) A formal process survey identifying specific dates and types of process and control changes. 3) An Historical Relative Exposure Rating Scale; plant health and safety personnel used this spreadsheet-based rating scheme to quantify exposures before and between years of actual measurement relative to a reference year in which measurements were available. A job-exposure matrix was calculated by integrating these three methods. Linear scaling factors were identified to interconvert geometric to arithmetic means and to interconvert total and inhalable dust. Individual worker cumulative exposures were then calculated based upon job histories linked with the job-exposure matrix. The nine-step process for integrating all available relevant data was effective in estimating the exposures for each of the cells of the job-exposure matrix. Among the 1680 workers participating, the mean cumulative inhalable dust exposure was 48.4 mg-years/m(3). Early years contribute disproportionately to the cumulative exposures of individuals since levels have declined significantly over time. The use of multiple sources of information, including a relative exposure rating instrument, significantly facilitates reconstruction of historical exposures. Inadequate adjustment for temporal trends can lead to underestimation of cumulative exposures and significantly affect estimation of dose-effect relationships. These methods are applicable to other situations requiring estimation of cumulative exposure with sparse industrial hygiene data in early years.