MSc Stats Project Presentation "Joint Modelling of Progression Free Survival and Overall Survival" by Sarah Rathwell

SARAH RATHWELL

ABSTRACT

Progression free survival is often measured along with overall survival to analyze the efficacy of treatments, or the prognostic capabilities of disease markers. This project outlines two approaches, based on either exponential distributions or Weibull distributions, to jointly model overall and progression free survival. Each approach allows for estimation of marginal overall survival, as well as correlations between overall survival and disease progression. Simulated data and data from two longitudinal cohort studies are used to demonstrate and compare the methods. In all cases, the joint models returned similar survival estimates when compared to both a naive Weibull model for overall survival, and the Kaplan-Meier survival curve. Additionally, the joint models tend to mirror the results given by the original studies, and were able to give further insight into which events (progression, pre- or post-progression mortality) were most impacted by treatment or disease marker.