ABSTRACT
In environmental carcinogenesis testing, historical control data can add valuable information when assessing exposure-related tumor incidence with very rare cancers. If only a small number of tumors are observed in either the concurrent control or treatment group, a simple two-sample comparison may not possess sufficient power to detect a small but true tumorigenic effect. Use of appropriate historical control information in this case can increase the test’s power to detect the small effect. Historical information on environmental phenomena can be misused, and care must be taken when including historical data in any statistical analyses. For example, nomenclature conventions and diagnostic criteria must be calibrated carefully, so as to allow meaningful comparisons among similar effects. The historical data must be gathered by the same research organization that conducts or oversees the current study. One can also use historical data to test for increasing trend when an environmental agent is studied over a range of quantifiable measures, such as exposure doses.
