| Precision in measuring exposure is a fundamental element of epidemiological studies. The greater the precision of the exposure measurement, the greater the confidence with which findings are presented and interpreted. Accurate exposure assessment is key in determining the true effect of the exposure of interest, and the magnitude of that effect.; Blood samples were taken from 192 former workers at an electrical capacitor plant in central Illinois and 27 randomly chosen members of the same community. The serum concentrations of polychlorinated biphenyls (PCBs) were measured and all members of the cohort answered a two-part survey covering occupational and non-occupational exposures. The results were analyzed to determine how well job history variables would predict serum PCB levels and whether these variables would be better predictors of the blood levels expressed as one surnmary value, total PCBs, or as a number of well-defined subvalues, called PCB congener groups. Individual PCB congeners and congener groups were also analyzed as proportions of total PCBs to look at congener distribution and sample composition.; Correlation and multiple regression analyses demonstrated that work variables could be useful in some settings to predict exposure levels, and that meaningful differences exist between total PCBs and congener-specific as exposure measures. Several regression models using a time of employment variable with gender and job title as predictors resulted in a fit with R2 = >0.60.; These findings have important implications not only for both exposure assessment and methodology, but for policy and regulatory applications as well. |
