The goal of this project is to develop improved statistical methods for toxicology studies. Work has proceeded in two areas: (1) the analysis of tumor multiplicity data, and (2) risk assessment and testing in reproductive toxicity studies that measure multiple endpoints. In tumorigenicity experiments that utilize animal models of skin and breast cancer, the tumor burden on each animal is evident and can be measured at repeated observation times. These types of studies are widely used to explore cancer mechanisms, and to identify carcinogens and compounds with potential chemopreventive attributes. We have developed five new models for analyzing tumor burden data: (1) one that assumes tumors are irreversible, (2) one that allows global and separate testing of effects on tumor onset, multiplicity, and regression, (3) one that enables testing for global effects in sparse data sets, (4) one that allows separation of effects on the number of initiated cells and the tumor latency, and (5) one that incorporates biologically reasonable assumptions to improve efficiency. Multiple endpoints are often measured in reproductive toxicity studies. I have developed two methods for assessing overall toxic effects when endpoints include both the number of subunits per dam (litter size, number of implants) and multiple binary outcomes on each subunit (implant resorbed, fetus malformed). I have also developed a general framework for modeling multivariate clustered data that enables joint estimation of effects on disparate outcomes such as the number of implantation sites per animal, the proportion of dead fetuses per dam, the proportion of malformed fetuses per dam, and birth weight.
Showing the most recent 10 out of 18 publications
