Abstract

OBJECTIVES: The aim of this study was to assess the theoretical efficiencies of different grouping strategies and its effect on the exposure-response relation in a study of respiratory morbidity associated with exposure to total inhalable and respirable carbon black dust. METHODS: A large epidemiological study is being undertaken to investigate the respiratory health of employees in the European carbon black manufacturing industry in relation to exposure to carbon black dust. In phase 2 of the study, repeated measurements of total inhalable and respirable dust were taken which enabled estimation of various components of variability in the exposure data (within and between worker variance and within and between group variance). These variance components were used to calculate the contrast in exposure between the groups in various classification schemes and to calculate the theoretical attenuation of the exposure-response relation and the standard error (SE) of the slope. RESULTS: High contrast in exposure was found when workers were classified according to the combination of their factory and job category as well as when these combinations were amalgamated into five exposure groups. Attenuation was minimal with most grouping schemes; only with the individual based strategy was the attenuation large. The SE of the theoretically attenuated exposure-response slope was smallest for the strategy based on individual people followed by the classification scheme based on factory and job category. CONCLUSIONS: It was concluded that, although some assumptions for the calculations of the attenuation of the exposure-response slope were not met, the most appropriate classification scheme of the worker seems to be by the combination of factory and job category.