B0902
Title: Causal effects on time-to-event outcomes in an oncology RCT with treatment discontinuation
Authors: Veronica Ballerini - University of Florence (Italy) [presenting]
Abstract: In clinical trials, patients sometimes discontinue study treatments prematurely due to reasons such as adverse events. Since treatment discontinuation occurs after the randomisation as an intercurrent event, it makes causal inference more challenging. The intention-to-treat analysis provides valid causal estimates of the effect of treatment assignment; still, it does not take into account whether or not patients had to discontinue the treatment prematurely. The problem of treatment discontinuation is proposed to deal with using principal stratification. Under this approach, the overall ITT effect is decomposed into an infinite number of principal causal effects for groups of patients defined by their potential discontinuation behaviour in continuous time. A flexible model-based Bayesian approach is used for inference, taking into account that discontinuation happens in continuous time, discontinuation time is not defined for patients who would never discontinue, and time to progression or death and discontinuation time are subject to administrative censoring. The framework is applied to analyse synthetic data based on a recent clinical trial in oncology, aiming to assess the causal effects of a new investigational drug combined with standard of care versus standard of care alone on progression-free survival. Finally, it is highlighted how such an approach makes it straightforward to characterise patients' discontinuation behaviour with respect to the available covariates.