censoring when simulating data

[adibender]

One thing we (not necessarily @pkopper) should think about in the context of #106 is censoring.

So far I always added censoring manually after the call to sim_pexp. The question is if that could be usefully integrated within sim_pexp, potentially as a special case of CR (where two linear predictors/event times are generated but independent of each other and the smaller one is selected) or if its better to add a function add_censoring that can be called after generating the survival times, e.g.

The former would require some way to specify if rhs_censoring is a competing risk or a censoring specification, for which I don't see an easy way to do...

data %>% sim_pexp(~rhs_event|rhs_censoring)
# or
data %>% sim_pexp(~rhs) %>% add_censoring(~rhs_censoring)

[pkopper]

Thank you for that note.
I tend to the first suggestion of your intuitively as this is as I would implement it in the non-PAMM set up, too.
However, I'll have a look at it and assign myself.

[adibender]

this will go into https://github.com/adibender/simpexp