Issues in performing retrospective exposure assessment.
Appl Occup Environ Hyg 1991 Jun; 6(6):421-427
A discussion was presented of limitations of traditional retrospective exposure measures and benefits of quantitative exposure assessments. Use of a subject's employment within an industry or job as a surrogate for exposure was compared with use of specific exposures. Analysis of relative risks of laryngeal cancer based on employment in a specific industry yielded lower risks than those produced by analysis by potential exposures to specific chemicals. Only the latter produced statistically significant associations. Analysis by job titles gave similar estimates as those from exposure analyses, but larger confidence intervals were obtained in the former. Job title analyses did not directly relate to specific types of exposures. Dose response relationships could be studied using duration of employment or exposure as surrogates for actual exposure data only when certain conditions were met. Analysis of exposure monitoring data obtained from 1930 to 1975 in a chrysotile (12001295) asbestos facility demonstrated that exposure levels varied between time periods and among departments. Thus, inappropriate grouping by duration of employment or exposure could produce serious misclassification of exposures. A study of formaldehyde (50000) exposure data revealed a poor association of employment duration with estimated exposures. Analysis of data from 25 studies indicated better exposure response relationships for cumulative exposure and exposure intensity compared with exposure duration. A semiquantitative approach based on relative exposure categories arbitrarily assigned weights to each category, which could cause misclassifications. Development of quantitative estimates and use of these to classify subjects by exposure categories could reduce potential exposure misclassifications. The authors conclude that use of the most quantitative procedure possible for exposure estimates reflective of dose will increase the power of epidemiologic studies of exposure response relationships.
NIOSH-Publication; NIOSH-Author; NIOSH-Contract; Contract-91-38661; Humans; Occupational-health; Occupational-exposure; Exposure-levels; Epidemiology; Quantitative-analysis; Risk-analysis; Statistical-analysis; Industrial-exposures
Applied Occupational and Environmental Hygiene