Skip to content
Back to formatted view

Raw Message

Message-ID: <52AB06F4.9080204@mayo.edu>
Date: 2013-12-13T13:09:08Z
From: Terry Therneau
Subject: Coxme time dependent
In-Reply-To: <mailman.23.1386932407.8862.r-help@r-project.org>

On 12/13/2013 05:00 AM, r-help-request at r-project.org wrote:
> i all,
>
>
>
> I am using the coxme function to fit random effects survival models.  The
> base survival models show a typical "J" shaped curve where survival drops
> sharply then plateaus over time interval.  I want to include a time
> covariate which could account for this non-linear decrease in survival over
> time.  However, I have not found a clear way to do this in the coxme
> models.  Any suggestions would be greatly appreciated.  Below is some
> sample code and data.
>
>
>
> -Jared
>

I don't understand what you need.  One of the advantages of the Cox model is that the 
baseline hazard function automatically accommodates such features, for both coxph and coxme.

Terry Therneau

PS: I can't make out the pattern of your sample data.  Perhaps you were depending on the 
html formatting?  R-help uses simple text only.