[R] Time-dependent covariates in survreg function

[Dimitris Rizopoulos]

Dear Michael,

AFAIK survreg() cannot handle time-dependent covariates. In particular, 
things are getting more complicated under the accelerated failure time 
framework when it comes to the handling of time-dependent covariates.

Moreover, it is important (for both Cox and AFT models) to determine 
what type of the time-dependent covariate you have in your problem. Is 
it exogenous (aka external) or endogenous (aka  internal) -- for info 
have a look at [1, Section 6.3]. For exogenous time-dependent covariates 
you can still use the time-dependent version of the Cox model. However, 
for endogenous ones you might have considerable bias if you do so.

For endogenous time-dependent covariates an alternative is to use 
package JM (http://www.jstatsoft.org/v35/i09, 
http://CRAN.R-project.org/package=JM, 
http://rwiki.sciviews.org/doku.php?id=packages:cran:jm) that has both 
relative risk and AFT models available.

I hope it helps.

Best,
Dimitris


[1] Kalbfleisch, J. and Prentice, R. (2002). The Statistical Analysis of 
Failure Time Data, 2nd edition. John Wiley & Sons, New York.




On 7/28/2010 10:06 AM, Michael Haenlein wrote:
> Dear all,
>
> I'm asking this question again as I didn't get a reply last time:
>
> I'm doing a survival analysis with time-dependent covariates. Until now,
> I have used a simple Cox model for this, specifically the coxph function
> from the survival library. Now, I would like to try out an accelerated
> failure time model with a parametric specification as implemented for
> example in the survreg function.
>
> Two questions: First, can survreg handle time-dependent covariates?
> The description for this function does not make reference to them. And
> second, in case survreg cannot deal with time-dependent covariates, is there
> a similar function in some other package that can?
>
> Thanks very much,
>
> Michael
>
>
> Michael Haenlein
> Associate Professor of Marketing
> ESCP Europe
> Paris, France
>
> 	[[alternative HTML version deleted]]
>
> ______________________________________________
> R-help at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.
>

-- 
Dimitris Rizopoulos
Assistant Professor
Department of Biostatistics
Erasmus University Medical Center

Address: PO Box 2040, 3000 CA Rotterdam, the Netherlands
Tel: +31/(0)10/7043478
Fax: +31/(0)10/7043014