In the primary simulations, we set the cumulative disease incidence to be either 5, 10 or 15% regardless of the censoring distribution, as model choice is often solely based on the proportion of cases attained over follow-up. However, keeping the proportion of cases constant across censoring strategies changes the disease incidence rate. Hence, we performed a secondary set of simulations where the survey and random censoring strategies were applied to the event time distributions from the complete follow-up model.
