Time-varying covariates in the interval-censored Cox model 

Answer: If you are working with multiple-record-per-subject interval-censored data, your data would need to be in the long form to use *stintcox* . For these data, time-varying covariates can be included by using either the *tvc()* option or having some variables vary with time; also, you will use the *id()* option to identify observations belonging to the same subject. If you are working with single-record-per-subject interval-censored data, your data do not need to be in the long form. For these data, time-varying covariates can be included by using the *tvc()* option. Also, you will only have one observation per subject. However, note that the lower and upper endpoints for the interval should be stored in two separate variables in this case.