Up to now, there is no quick way to introduce time-varying covariates other than playing with interruption events.
I suggest to rename the prevous constructor'TimeVaryingCovariate' into 'EventCovariate'.
This type of covariate can be accessed and modified in CAMPSIS events (like previously)
I suggest then to reuse the constructor 'TimeVaryingCovariate' for classical time-varying covariates, with the following arguments:
name: name of the covariate
table: data.frame, must contain the mandatory columns 'TIME' and 'VALUE'. An 'ID' column may also be specified. In that case, ID's between 1 and the max number of subjects in the dataset/arm can be used. All ID's must have a VALUE defined for TIME 0.
This new type of covariate will be implemented using EVID=2 rows.
Up to now, there is no quick way to introduce time-varying covariates other than playing with interruption events.
I suggest to rename the prevous constructor'TimeVaryingCovariate' into 'EventCovariate'. This type of covariate can be accessed and modified in CAMPSIS events (like previously)
I suggest then to reuse the constructor 'TimeVaryingCovariate' for classical time-varying covariates, with the following arguments: name: name of the covariate table: data.frame, must contain the mandatory columns 'TIME' and 'VALUE'. An 'ID' column may also be specified. In that case, ID's between 1 and the max number of subjects in the dataset/arm can be used. All ID's must have a VALUE defined for TIME 0.
This new type of covariate will be implemented using EVID=2 rows.