[R] ID parameter in model

[Göran Broström]

On Mon, May 2, 2011 at 5:38 PM, Mike Harwood <harwood262 at gmail.com> wrote:
> Hello,
>
> I am apparently confused about the use of an id parameter for an event
> history/survival model, and why the EHA documentation for aftreg does
> not specify one.  All assistance and insights are appreciated.

Which version of eha are you using? The latest version documents the
use of 'id'.

> Attempting to specifiy an id variable with the documentation example
> generates an "overlapping intervals" error,

Thanks for pointing this out. It is an error (actually three) in the
data frame.
The reason is that it is "real" data; and not sufficiently checked by me.

I'll fix this in an update soon.

Thanks again,

Göran

 so I sorted the original
> mort dataframe and set subsequent entry times an id to the previous
> exit time + 0.0001.  This allowed me to see the affect of the id
> parameter on the coefficients and significance tests, and prompted my
> question.  The code I used is shown below, with the results at the
> bottom.  Thanks in advance!
>
> Mike
>
> head(mort) ## data clearly contains multiple entries for some of the
> dataframe ids
>
> no.id.aft <- aftreg(Surv(enter, exit, event) ~ ses, data = mort)  ##
> Inital model
> id.aft <- aftreg(Surv(enter, exit, event) ~ ses, data = mort, id=id)
> ## overlapping intervals error
>
> mort.sort <- ## ensure records ordered
>    mort[
>        order(mort$id, mort$enter),]
>
> ## remove overlap
> for (i in 2:nrow(mort.sort)){
>     if (mort.sort[i,'id'] == mort.sort[i-1,'id'])
>         mort.sort[i,'enter'] <- mort.sort[i-1, 'exit'] + 0.0001
>        }
>
> no.id.aft.sort <- aftreg(Surv(enter, exit, event) ~ ses, data =
> mort.sort) ## initial model on modified df
> id.aft.sort <- aftreg(Surv(enter, exit, event) ~ ses, id=id, data =
> mort.sort) ## with id parameter
>
>
> #=== output ===========#
>> no.id.aft.sort
> Call:
> aftreg(formula = Surv(enter, exit, event) ~ ses, data = mort.sort)
>
> Covariate          W.mean      Coef Exp(Coef)  se(Coef)    Wald p
> ses
>           lower    0.416     0         1           (reference)
>           upper    0.584    -0.347     0.707     0.089     0.000
>
> log(scale)                    3.603    36.704     0.065     0.000
> log(shape)                    0.331     1.393     0.058     0.000
>
> Events                    276
> Total time at risk         17045
> Max. log. likelihood      -1391.4
> LR test statistic         16.1
> Degrees of freedom        1
> Overall p-value           6.04394e-05
>> id.aft.sort
> Call:
> aftreg(formula = Surv(enter, exit, event) ~ ses, data = mort.sort,
>    id = id)
>
> Covariate          W.mean      Coef Exp(Coef)  se(Coef)    Wald p
> ses
>           lower    0.416     0         1           (reference)
>           upper    0.584    -0.364     0.695     0.090     0.000
>
> log(scale)                    3.588    36.171     0.065     0.000
> log(shape)                    0.338     1.402     0.058     0.000
>
> Events                    276
> Total time at risk         17045
> Max. log. likelihood      -1390.8
> LR test statistic         17.2
> Degrees of freedom        1
> Overall p-value           3.3091e-05
>>
>
> ______________________________________________
> 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.
>



-- 
Göran Broström