As an applied statistician, my work mixes methodological development, software development, and data analysis. By methodological development I mean adapt or combine existing existing statistical results to solve a statistical issue. So far, I have organized my research around three topics:

  • the development of models for repeated measurements and complex systems. This includes latent variable models and linear mixed models. I apply these developments on neuroimaging data and data from psychological tests to investigate the relationship between the brain response and depression.

  • the analysis of registry data in presence of right-censoring, competing risks, and counfounding. The aim is to develop estimators that are robust (to model misspecification) and efficient based on the semi-parametric theory. My developments have, for instance, been applied to compare preventive stroke treatments using the danish registry data.

  • the development of generalized pairwise comparisons to handle multiple and heterogenous outcomes. A typical application is the assessment of new chemotherapy where being able to balance the benefit (longer in survival time) and the risk (treatment side effects) in an interpretable way is important.

I have implemented these developments in the R packages lavaSearch2, LMMstar, riskRegression, and BuyseTest.

I have been also been reading about:

  • Group sequential design with Group Sequential and Confirmatory Adaptive Designs in Clinical Trials (Wassmer, 2006) and Group Sequential Methods with Applications to Clinical Trials (Jennison and Turnbull, 1999)
  • Causal inference with Causal Inference: What If (Hernan, Robins, 2020), Causal Inference in Statistics - A Primer (Pearl, 2016), and Causality: Models, Reasoning and Inference (Pearl, 2000)
  • Semi parametric theory with Semiparametric Theory and Missing Data (Tsiatis, 2006)
  • Statistical inference, especially multiple testing with Multiple comparisons using R (Bretz, 2011) and non-parametric tests with Rank and Pseudo-Rank Procedures for Independent Observations in Factorial (Brunner 2018)
  • Splines with the excellent book Generalized Additive Models: An Introduction with R (Wood 2017)
  • Asymptotic theory with Asymptotic statistics (van der Vaart, 1998), In all likelihood (Pawitan, 2001), and U-Statistics: theory and practice (Lee, 1990). I have also read most of High-Dimensional Statistics: A Non-Asymptotic Viewpoint (Wainwright., 2019) but would not claim to master the topic.
  • Epidemiology with the classical reference Statistical Methods in Epidemiology (Clayton and Hills, 1993), a nice and concise book.
  • Time to event analysis with the nice book Dynamic Regression Models for Survival Data (Martinussen and Scheike, 2006).
  • Post-selection inference with mostly papers like Exact post-selection inference, with application to the lasso (Lee et al., 2016) or Optimal Inference After Model Selection (Fithian et al., 2014)
  • Communication in science with Designing Science Presentations (Carter, 2021)

I’m currently reading about:

  • Dynamic Treatment Regimes with Dynamic Treatment Regimes: Statistical Methods for Precision Medicine (Tsiatis et al., 2019)
  • Survival analys with Survival and event history analysis: a process point of view (Aalen, Borgan, and Gjessing, 2008) and Counting processes and Survival analysis (Fleming and Harrington, 2005)