Instead, exposure assessments are based on chemical measurements in either environmental media or biological specimens. In either case, exposure serves as a surrogate for dose. Exposure assessments based on measurements of chemical contaminants in the environment attempt to quantify the amount of the contaminant that contacts a body barrier over a defined time period. Exposure can occur via three routes: inhalation, skin contact, and ingestion. Exposure can also be assessed by measuring chemicals or their metabolites in human tissues. Such biomarkers of exposure integrate absorption from all routes. The evaluation of biomarkers can be complex, since most markers are not stable for long periods of time. Knowledge of pharmacokinetics is essential to the linkage of measurements at the time of sampling with past exposures. Similarly, biomarkers that have the possibility of being biomarkers of effect, such as DNA adducts, show promise, but do not necessarily provide accurate measures of past exposures; that is, there is no evidence that currently measured DNA adducts have any relationship to occupational or environmental exposures experienced years before.
Quantitative assessments based on environmental or biologic samples are rarely available for epidemiologic studies; instead, investigators must rely on a mixture of qualitative and quantitative information to produce exposure estimates. One can usefully distinguish a few basic approaches to exposure assessment for epidemiology (Checkoway et al., 1989; Armstrong et al., 1994). The simplest approach compares the members of a group presumably exposed to a toxic agent with the general population or with a nonexposed group. The advantages of that approach are its simplicity and the ease of interpretation of results. If, however, only a small fraction of the group is exposed to the agent, any increased risk posed by exposure of this subgroup may not be detectable when the risk of the entire group is assessed.
A more refined method of exposure assessment assigns each study subject to an exposure category, such as high, medium, low, and no exposure. Disease risk in each group can then be calculated separately and compared with a reference or nonexposed group. This method, in contrast with the simple exposed–nonexposed comparison above, can evaluate the presence or absence of a dose–response trend. In some cases, more detailed information is available, and quantitative exposure estimates can be developed. Such estimates are sometimes referred to as exposure metrics. Exposure metrics integrate quantitative estimates of exposure intensity (such as air concentration or extent of skin contact) with exposure duration to produce an estimate of cumulative exposure. Ideally, such refined estimates reduce errors associated with misclassification and thereby increase the power of statistical analysis to identify true associations between exposure and disease.
Occupational-exposure studies tend to rely on work histories, job titles, and workplace measurements of contaminant concentration, which can be combined to create a job–exposure matrix, in which a quantitative exposure estimate is