create a hazard function for censoring time AND a hazard function for survival time, then create the observed duration as min(survival, censoring). See the second paragraph on page 1 here. I think we would want to use another draw from the same hazard function to do that. That would give us non-informative censoring from conditional independence of the survival time and censoring time, given the covariates.
create a hazard function for censoring time AND a hazard function for survival time, then create the observed duration as min(survival, censoring). See the second paragraph on page 1 here. I think we would want to use another draw from the same hazard function to do that. That would give us non-informative censoring from conditional independence of the survival time and censoring time, given the covariates.