Anna Decker

Biostatistics Ph.D. Program
B.S. Mathematical Biology Beloit College, 2009
M.A. Biostatistics, UC Berkeley, 2011
Office / Location: 
439 Evans Hall
Email: 
decker [at] stat [dot] berkeley [dot] edu
Primary Research Area: 
Applied & Theoretical Statistics
Sub-Focus: 
Bioinformatics/Biostatistics, Machine Learning
Dissertation: 
Semiparametric, prediction, variable importance, and longitudinal causal effect estimation in critical care
Dissertation Advisor: 
Alan Hubbard

I am a fifth year PhD candidate in Biostatistics. My research is focused on the application of causal inference and machine learning in the study of severe trauma. In collaboration with clinicians at San Francisco General Hospital, we have developed time-specific prediction models for mortality and blood product infusion and determine rankings of clinical variables at each time point based on parameters motivated by the causal inference literature. I am also interested in the longitudinal dynamic allocation of blood products based on patients' health characteristics at a given time point. In addition to my work in trauma injury, I collaborate with the division of Environmental Health in the School of Public Health examining the mediation of genetic effects by epigenetic factors such as methylation and their effects on cancer risk. I will graduate in May 2014 and am interested in continuing to work at the intersection between statistics and medicine/public health.