[R] exclude missing co-variable data in cox model

Therneau, Terry M., Ph.D. therneau at mayo.edu
Fri Dec 19 13:58:29 CET 2014

Three responses to your question
   1. Missing values in R are denoted by "NA".  When reading in your data you want to use 
the "na.strings" option so that the internal form of the data has missing values properly 

   2. If this is done, then coxme will notice the missings and remove them, you do not 
need to do anything.  Your second question of "how to use the missing data" is a much 
deeper statistical one.  Multiple imputation would be the obvious way to proceed, but it 
is complex and I have no experience with respect to how it would work with a mixed effects 
Cox model.

   3. Your note implies that output was attached.  Note that r-help strips away all 
attachments, so none of us saw it.

Terry Therneau

On 12/19/2014 05:00 AM, r-help-request at r-project.org wrote:
> Hi all,
> I have a data set like this:
> Test.cox file:
> V1        V2         V3                Survival       Event
> ann      13          WTHomo          4                1
> ben      20          *                        5                1
> tom      40         Variant               6                1
> where "*" indicates that I don't know what the value is for V3 for Ben.
>  Remainder of message excluded

More information about the R-help mailing list