[R] coxph and survfit issue - strata

David Winsemius dwinsemius at comcast.net
Sat May 7 04:02:52 CEST 2011


On May 6, 2011, at 6:22 PM, Eva Bouguen wrote:

> Dear users,
>
> In a study with recurrent events:
> My objective is to get estimates of survival (obtained through a Cox  
> model) by rank of recurrence and by treatment group.
> With the following code (corresponding to a model with a global  
> effect of the treatment=rx), I get no error and manage to obtain  
> what I want :
> data<-(bladder)
> model1<-coxph(Surv(stop,event)~rx+strata(enum) 
> +cluster(id),data=bladder)
> data1<-data.frame(rx=1)
> survfit(model1,newdata=data1)
>
> But with the model with strata by treatment interaction  
> (corresponding to a model with an effect of the treatment by rank),  
> I get the following error:
> model2<-coxph(Surv(stop,event)~rx*strata(enum) 
> +cluster(id),data=bladder)
>
> data1<-data.frame(rx=1)
>
> survfit(model2,newdata=data1)
> Erreur dans strata(enum) : objet "enum" non trouvé =>error in  
> strata(enum) : object "enum" not found

Right. Looks like a perfectly clear error message to me .... you  
created a newdata dataframe that did not have all the necessary  
columns that you used on the RHS of your formula.

-- 

David Winsemius, MD
West Hartford, CT



More information about the R-help mailing list