Where am I?
An important part of the SBIM's activity is devoted to methodological research. Different axes are explored in biostatistical research, the main ones being:
- Bayesian inference
Our goals are to propose alternative methods for conducting clinical trials by proposing approaches that better manage current knowledge about the efficacy or toxicity of a new drug. These methods extend to the evaluation of new diagnostic tests as well as quality control.
- Causal inference
We develop approaches for the analysis of data from non-interventional clinical studies, taking into account the specificities of sample size, prevalence of the treatment group, and the binary or censored nature of the criteria.
- Prognostic models
Our objectives are to provide measures to assess the prognostic value of a new classification or prognostic marker, to take into account missing data in such modelling, and to study different strategies for joint modelling of longitudinal and survival data (in the broad sense). In particular, we develop multiple imputation techniques to process missing data with a particular focus on survival data and clustered data.