Abstract: Beginning with my time as a student of statistics in Michigan, to obtaining a Ph.D. in Biostatistics at Harvard, to working as a statistician in the pharmaceutical industry in Switzerland, I will first share my career journey. Next, I will discuss my current role at Roche as a statistical methodologist supporting a wide variety of clinical trials for new treatments in a broad range of diseases from Parkinson’s Disease to Cancer. Pharmaceutical statistics remains a stronghold of classical statistical inference, due to the fact that regulatory bodies such as the FDA and EMA require strict type-I error control and that statistical power to reject the null hypothesis lies at the heart of a well-designed clinical trial. 

In the second half of the talk, I’ll describe my current area of statistical research in causal mediation analysis. I’ll motivate the relevance of using mediation analysis to elucidate causal pathways for new treatments and present a novel method for mediation for time-to-event endpoints using pseudo-values with an application to a multiple sclerosis clinical trial. In the process, I will touch on challenges in simulation studies for mediation and our proposal to apply Gaussian quadrature to compute the target estimands since true values of interest are not explicit parameters of the data-generating mechanism. Extensions of these mediation methods are the subject of an open internship for which interested students are encouraged to apply. 

About the Speaker: Alex Ocampo is a Statistical Methodologist with Roche based in Basel, Switzerland. He obtained his Bachelor’s degree in Statistics from the University of Michigan and Ph.D. in Biostatistics from Harvard University in 2020 where his dissertation focused on statistical methods for dealing with missing data. In his current work, he supports drug development teams at Roche in implementing the most impactful statistical methodologies in their clinical trials. More broadly, he focuses on promoting causal thinking in the pharmaceutical industry and co-leads a cross-industry working group on causal inference and a sub-team of this group on mediation analysis.

His current statistical research focuses on causal inference, mediation analysis, surrogacy, Bayesian methods, semiparametric theory, and multi-component endpoint analyses.