Occupational exposure assessment in epidemiology - principles in the design of an optimal exposure assessment strategy.
Heederik-D; Kromhout-H; Attfield-M
Proceedings of the 9th International Symposium on Epidemiology in Occupational Health: Book of Extended Abstracts from the Proceedings of the 9th International Symposium on Epidemiology in Occupational Health, September 23-25, 1992, Cincinnati, Ohio. Cincinnati, OH: U.S. Deptartment of Health and Human Services, Public Health Service, Centers for Disease Control and Prevention, National Institute for Occupational Safety and Health, DHHS (NIOSH) Publication 94-112, 1994 Jan; :29-55
Principles used in the design of an optimal exposure assessment strategy in occupational exposure studies were investigated. Approaches used in the assessment of occupational exposures in epidemiological studies were job exposure matrices, toxicokinetic modeling, and time frame modeling. No clear guidelines existed, however, as to which index of dose was optimal in a given situation. The influence of misclassification of exposure was addressed. Misclassification introduced a bias in the estimate of the magnitude of the relationship between exposure and effect. The lack of qualitative information on exposure also led to the employment of methods such as questionnaires, although the validity and reproducibility of information elicited through them was known to be limited. Methods of semiquantitative measure using a ranking scale based on tasks performed and the nature of the work process experienced problems when it came to ranking exposures across different industries. Changes in the cutoff points for considering the high and low exposure levels changed the sensitivity and specificity of the exposure matrices, and impacted on risk ratios. Quantitative exposure assessments of exposure were usually considered to be superior to other methods of estimation. Estimates based on individual exposures were acceptable only if repeated measurements were made in order to diminish day to day variability. Quantitative measurements using groups were discussed in relation to intragroup variance (homogeneity), intergroup variance (contrast), and the number of measurements in a category (precision). The concept was illustrated by a reanalysis of data from Dutch grain processing and animal feed workers. Data from grouping strategies were less prone to attenuation effects than those from individuals. The authors conclude that hierarchies in exposure measures are a simplification of the real situation, and that more emphasis on determining the precision and validity of exposure measures is needed.
NIOSH-Author; Air-contamination; Environmental-pollution; Epidemiology; Lung-disorders; Mathematical-models; Occupational-exposure; Occupational-hazards; Toxicopathology
Proceedings of the 9th International Symposium on Epidemiology in Occupational Health: Book of Extended Abstracts from the Proceedings of the 9th International Symposium on Epidemiology in Occupational Health, September 23-25, 1992, Cincinnati, Ohio