Correcting for Non-Compliance in Randomized Trials
*James Robins, Harvard School fo Public Health 


This presentation will be based on the papers published by Dr Robins et al in Comm, Stat, Theory Meth (1991) and in Biometrics (2000). The inverse probability of censoring weighting (IPCW) method which uses weighting approach to adjusted based sampling and the rank preserved structural failure time (RPSFT) model which models failure time scale itself, will be illustrated to address issues of non-compliance and treatment cross-over in clinical trials.