Occupational studies have identified many of the established chemical carci
nogens. Studies in the next millennium will be needed to identify the hazar
dous agents in occupations known to have high cancer rates, to assess human
risks from animal carcinogens that have not been well evaluated epidemiolo
gically, to provide information on women and minorities, to evaluate intera
ctions with genetic factors and other risk factors, to contribute to our un
derstanding of risks from the spread of chemicals from the workplace to the
general environment, and to identify mechanisms of cancer. The traditional
retrospective cohort design will be insufficient to meet these needs. Popu
lation-based case-control, nested case-control, prospective cohorts, and cr
oss-sectional designs will assume more important roles because of the need
to collect information on nonoccupational risk factors and biological tissu
es. Improvement in the assessment of quantitative exposures is needed for t
he efficient evaluation of interactions between occupational exposures, gen
etic factors, and nonoccupational exposures.