This project is concerned with statistical methodology issues in the design, analysis, and interpretation of laboratory animal experiments. Specific research efforts include: (1) It was shown that background tumor rates in F344 rats a much higher in current studies than in those conducted a decade ago. The increased tumor rates are accompanied by decreased survival and increased body weights. Factors responsible for these increasing tumor rates include more tissue examined at necropsy in current studies, changes in histopathology diagnosis genetic drift, and dietary factors. (2) Methodology was developed to permit a survival-adjusted analysis of tumor incidence rates while avoiding the need to make assumptions about tumor lethality or cause of death. If constraints are placed on the transition rates describing the underlying tumor onset/death process, the analysis is feasible with only one sacrifice time (e.g., the terminal kill). The most promising restriction was the one that specified a constant difference between the death rates for animals with and without the tumor of interest. (3) Some of the statistical properties of computerized measurements sperm parameters such as velocity were examined. The analysis of such data complicated by having multiple sperm from the same animal, giving the nested structure of sperm within animal within treatment. Future research related to this project includes examining the relationship between maximum tolerated dose and rodent carcinogenicity, and comparing by computer's simulation existing methods with the incidence analysis based on constant differences between the death rates for tumor-free and tumor-bearing animals.
