Use of historical controls for animal experiments

Area of Science:

• Biostatistics
• Toxicology
• Carcinogenicity Bioassays

Background:

• Historical control data is crucial for interpreting rodent bioassays.
• Existing statistical methods for trend testing need refinement.
• Accurate trend analysis is vital for carcinogenicity assessment.

Purpose of the Study:

• To review and develop statistical methods for using historical control data.
• To compare the power of the Hoel-Yanagawa and Tarone tests.
• To assess the fit of beta-binomial distributions to historical data.

Main Methods:

• Development of asymptotic properties for Hoel-Yanagawa exact conditional tests.
• Comparison of Hoel-Yanagawa tests with the Tarone test.
• Estimation of beta-binomial parameters from historical control data.
• Application to the National Toxicology Program historical control database.

Main Results:

• The Hoel-Yanagawa test is found to be more powerful than the Tarone test.
• Beta-binomial distributions show good fit to historical control data.
• The study illustrates the application of these methods to real-world data.

Conclusions:

• The Hoel-Yanagawa test offers improved power for trend detection in carcinogenicity studies.
• Beta-binomial models provide a robust framework for incorporating historical data.
• Consideration of test sensitivities and conservative usage is recommended.