A fitted model claims to be of class coxph, yet using terms(fit), which is defined for objects of class coxph, results in Error in terms.default(fit) : no terms component nor attribute. This, among other things, impedes using survminer::ggadjustedcurves to control for covariates (side note: would this approach actually be valid for Firth's penalization?). Thanks in advance :-)
Hi, thanks fo raising this issue! The fitted model is its own class coxphf which inherits some features from coxph. In #9, I added terms to the returned object, so terms(fit) should work now.
A fitted model claims to be of class
coxph
, yet usingterms(fit)
, which is defined for objects of classcoxph
, results inError in terms.default(fit) : no terms component nor attribute
. This, among other things, impedes usingsurvminer::ggadjustedcurves
to control for covariates (side note: would this approach actually be valid for Firth's penalization?). Thanks in advance :-)