Liangyuan Hu, Biostatistics Doctoral Graduate Program
Liangyuan Hu, PhD candidate in biostatistics, recently successfully defended her dissertation, titled “Causal Models for Comparing and Optimizing HIV Treatment Strategies Using EHR Data.” A synthesis of computer science, medicine, and statistics, her dissertation demonstrates the multidisciplinarity often involved in public health research. Hu used data collected from electronic health records (EHR) to determine how soon after diagnosis patients with infectious diseases such as HIV and tuberculosis should begin treatment. Patients receiving treatment too early are at risk for drug toxicity, while those receiving it too late face increased rates of morbidity and mortality. Using advanced statistical analyses and computer modeling, Hu identified optimal times for treatment commencement and duration.
Below are Hu’s answers to questions about her experience as a doctoral student in biostatistics.
Tell us about your time at Brown. What did you enjoy most about being a doctoral student here?
Brown is small but perfectly formed and equipped with diverse resources. Faculty at the School of Public Health have a wide range of expertise. Graduate students are exposed to various projects before they decide on a topic for their dissertation research. I began to work with Professor Joe Hogan three years ago as a research assistant. He introduced me to various projects and I decided that I was most interested in causal inference models in comparing and optimizing HIV treatment strategies. When mentoring me, he always pointed me to the right direction, but allowed me enough time and space to conduct independent research. I felt I could benefit a great deal from his mentorship and asked if he'd like to be my adviser. I had a great and pleasant PhD journey working with Professor Hogan.
What public health issues do you hope to address through your dissertation?
My dissertation addresses statistical issues in estimating causal effects of time-dependent treatment rules and comparing efficacy of treatment rules. The research methods developed in my dissertation are motivated by electronic health records (EHR) data on HIV treatment rules. EHR data provide a wealth of opportunities for care and research but because they are ‘in the field’, complications must be addressed before drawing valid inference. I hope that my dissertation research adds to the current literature by developing models and methods and investigating the research questions in greater details; complimenting randomized control trial findings; and deriving evidence from ‘field data' to inform guidelines.
What are your future plans now that you have finished your PhD?
I will be joining Analysis Group (AG) as an associate. At AG – a consulting firm that specializes in economics, finance, and health care analytics – I will continue to be a biostatistician working towards addressing statistical issues in health outcomes, health care services, health policy, economics, etc. by developing methods and applying cutting-edge statistical methods and computation techniques in these research areas.