My overarching goal is to produce robust and accessible statistical methods, with an emphasis on tools for precision medicine. I work to develop new approaches defining optimal treatment sequences to address important clinical questions such as “What antiretroviral drug should be given to a patient with HIV who has been diagnosed with non-alcoholic fatty liver disease?” or “What is the best sequence of antidepressants, given a patient’s personal characteristics, health status and preferences?”
Clinical trials, which are the conventional mode of clinical investigation, are often unable to provide the data needed to study precision medicine and sequential treatments. In contrast, data from non-experimental sources such as clinical registries or electronic health records provide unprecedented opportunities to learn about treatment tailoring. Many methodological challenges must be overcome before the full power of these data sources can be fully exploited.