Covariate "time of entry in the state" in clock reset transitions

[caterinagregorio]

Hello,

According to the current implementation of the package, is it possible to consider that the transition hazard in a iCTSTMs under the "clock reset" time scale depends also on the time at which the subject entered the state?

Thank you!

[mclements]

To my knowledge, this is not currently implemented.

In principle, this is possible, but it would need a mechanism to pass two time scales to the time-to-event models.

Devin: how difficult would it be to implement the two time scales?

[dincerti]

The general solution is to allow for covariates that dynamically update during the individual-level simulation. You would ideally want users to be able to write custom functions to update the covariates; however, thats would require some refactoring because the simulation is written in C++ at the moment and passing an R function to each iteration of a C++ for loop is highly inefficient.

@mclements how were you envisioning the two time scales would work?