Search

The aim of this master thesis is to develop a method for modeling direct and indirect effects in causal mediation analysis in the context of survival analysis, but where the mediator is also a time-to-event variable. We will take an interventional approach by Didelez (2019) and Aalen et al. (2020) where exposure is separated into one component, which affects the mediator, and one component, which affects survival directly. This approach will lead us to the non-parametric g-formula, introduced by Robins (1986), which will be combined with the additive hazards model for a survival outcome and the pooled logistic regression model for the mediator. We will illustrate the method with a simulation study as well as a real-life application on some breast cancer data.

The aim of this master thesis is to develop a method for modeling direct and indirect effects in causal mediation analysis in the context of survival analysis, but where the mediator is also a time-to-event variable. We will take an interventional approach by Didelez (2019) and Aalen et al. (2020) where exposure is separated into one component, which affects the mediator, and one component, which affects survival directly. This approach will lead us to the non-parametric g-formula, introduced by Robins (1986), which will be combined with the additive hazards model for a survival outcome and the pooled logistic regression model for the mediator. We will illustrate the method with a simulation study as well as a real-life application on some breast cancer data.