Because resources are limited, scientists must decide out of thousands of candidates which substances to select for testing in animals or in human cells. The tests are costly and time-consuming: determining whether a chemical causes cancer in rats or mice can cost several million dollars and take several years to complete. Three factors generally guide the decision to test a substance.

The Number of People Exposed

We want to test those chemicals that affect a large number of people or those for which the exposure levels have been unusually high. Pesticides, for example, fit both categories. They potentially affect a large number of people because of trace amounts on foods and their use in or around the home, and in farming-related occupations, exposure levels are high.

Previous Data

This could be a report that in laboratory cells, a chemical causes alterations in human DNA. Or it could be a report that people exposed to a particular chemical in the workplace or at a specific geographical location have developed specific cancers at higher rates than expected. This kind of information provides important clues for decisions about animal testing. Even before any animal testing was done, studies on human and animal cells indicated dioxins and polycyclic aromatic hydrocarbons were suspected carcinogens.

Public Concern

Chromium and some pesticides are examples of a group of concerned citizens first bringing to the attention of public health officials concerns about chemicals. The National Toxicology Program has a Web site available to the public to suggest agents suspected of causing cancer: ntp-server.niehs.nih.gov; click on “ Nominations to the Testing Program.”

Similar considerations guide epidemiologists as to whether to begin large population studies. Some of these factors include

• Animal study data suggesting a cancer-exposure link (e.g., vinyl chloride) or a related agent that raises suspicion (e.g., acrylonitrile was studied because of its structural similarity to vinyl chloride).
• Results from other epidemiologic studies (large population studies) that suggest the need for additional studies
• An exposure’s biological mechanisms that suggest a possible link to cancer.
• Pockets of specific cancers that cluster in a particular town or place or unusual case reports.
• Cancer trends—that is, rates that change over time or with location.
• Changes in cancer rates within a population after migration to a new area.
• Introduction of a new exposure or technology for which epidemiologic data are needed, or an unusual exposure pattern needing evaluation.

Key Points

• Resources are limited. Scientists must decide out of thousands of candidates which substances should be selected for animals or human cell testing. Tests are costly and time-consuming.
  • Accordingly, we want to test
    • Those chemicals that affect a large number of people, or
    • Those for which the exposure levels have been unusually high, or
    • We might want to test both.
  • Previous data provide important clues for decisions about animal testing. For example
    • A chemical causes alterations in human DNA in laboratory cells, or
    • A report that people exposed to a particular chemical in the workplace, or
    • A report that people at a specific geographical location are getting cancer at higher rates than expected, or
    • A combination of all three.
  • On occasion, a group of citizens first bring their concerns about chemicals to the attention of public health officials.
• Similar considerations guide epidemiologists as to whether to begin large population studies.