David Lloyd
2007-Aug-06 12:50 UTC
[R] (Censboot, Z-score, Cox) How to use Z-score as the statistic within censboot?
Dear R Help list,
My question is regarding extracting the standard error or Z-score from a
cph or coxph call. My Cox model is: -
modz=cph(Surv(TSURV,STATUS)~RAGE+DAGE+REG_WTIME_M+CLD_ISCH+POLY_VS,
data=kidneyT,method="breslow", x=T, y=T)
I've used names(modz) but can't see anything that will let me extract
the Z scores for each coefficient or the standard errors in the same way
one can use coef (modz).
I need to get either the se or Zscore as I am hoping to be able to use
the Zscore within a statistic function used in censboot. I wish to find
the proportion of times each of the variables in the cox model are
significant out of the R=1000 bootstrap resamples.
My current function for us with the bootstrap only gives me the
coefficients: -
func=function(kidneyT,indices)
{
kidneyT=kidneyT[indices,]
modz=coxph(Surv(TSURV,STATUS)~RAGE+DAGE+REG_WTIME_M+CLD_ISCH+POLY_VS,dat
a=kidneyT,method="breslow")
coefficients(modz)
}
I would like to be able to modify the following code so modz.cens.boot$t
below gives me an array of the Z-scores for each variable and each
resample - not the coefficients
modz.cens.boot=censboot(kidneyT, func,R=1000,sim="ordinary")
modz.cens.boot
I hope this is clear enough? I'm still getting to terms with the
Bootstrap, having only just started trying to use it a few days ago.
Any help on this would clear a massive headache for me as I've been
going around in circles for hours and hours on this.
Thx
DaveL
Click to find information on your credit score and your credit report.
<http://tagline.bidsystem.com/fc/Ioyw36XIe1IYw28KqOSu66l6ED7AsGYWlz4EWyK
N5OTcqKjQM8gfiW/>
<span id=m2wTl><p><font face="Arial, Helvetica,
sans-serif" size="2"
style="font-size:13.5px">_______________________________________________________________<BR>Get
the Free email that has everyone talking at <a href=http://www.mail2world.com
target=new>http://www.mail2world.com</a><br> <font
color=#999999>Unlimited Email Storage – POP3 – Calendar
– SMS – Translator – Much
More!</font></font></span>
[[alternative HTML version deleted]]
Frank E Harrell Jr
2007-Aug-06 15:56 UTC
[R] (Censboot, Z-score, Cox) How to use Z-score as the statistic within censboot?
David Lloyd wrote:> Dear R Help list, > > My question is regarding extracting the standard error or Z-score from a > cph or coxph call. My Cox model is: - > > modz=cph(Surv(TSURV,STATUS)~RAGE+DAGE+REG_WTIME_M+CLD_ISCH+POLY_VS, > data=kidneyT,method="breslow", x=T, y=T)You are assuming linearity for all predictors, which is unlikely. When you expand predictors into multiple parameters to allow for nonlinearity, or when you have factors with >2 levels, individual P-values will be meaningless. You can get all Wald multiple d.f. test stats and P-values by saving the result of anova(modz) using anova.Design, which creates a matrix. Frank> > I've used names(modz) but can't see anything that will let me extract > the Z scores for each coefficient or the standard errors in the same way > one can use coef (modz). > > I need to get either the se or Zscore as I am hoping to be able to use > the Zscore within a statistic function used in censboot. I wish to find > the proportion of times each of the variables in the cox model are > significant out of the R=1000 bootstrap resamples. > > My current function for us with the bootstrap only gives me the > coefficients: - > > func=function(kidneyT,indices) > { > > kidneyT=kidneyT[indices,] > modz=coxph(Surv(TSURV,STATUS)~RAGE+DAGE+REG_WTIME_M+CLD_ISCH+POLY_VS,dat > a=kidneyT,method="breslow") > > coefficients(modz) > > } > > > I would like to be able to modify the following code so modz.cens.boot$t > below gives me an array of the Z-scores for each variable and each > resample - not the coefficients > > modz.cens.boot=censboot(kidneyT, func,R=1000,sim="ordinary") > modz.cens.boot > > > I hope this is clear enough? I'm still getting to terms with the > Bootstrap, having only just started trying to use it a few days ago. > > Any help on this would clear a massive headache for me as I've been > going around in circles for hours and hours on this. > > Thx > > DaveL > > > > > > Click to find information on your credit score and your credit report. > <http://tagline.bidsystem.com/fc/Ioyw36XIe1IYw28KqOSu66l6ED7AsGYWlz4EWyK > N5OTcqKjQM8gfiW/> > > > > <span id=m2wTl><p><font face="Arial, Helvetica, sans-serif" size="2" style="font-size:13.5px">_______________________________________________________________<BR>Get the Free email that has everyone talking at <a href=http://www.mail2world.com target=new>http://www.mail2world.com</a><br> <font color=#999999>Unlimited Email Storage – POP3 – Calendar – SMS – Translator – Much More!</font></font></span> > [[alternative HTML version deleted]] > > ______________________________________________ > R-help at stat.math.ethz.ch 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. >-- Frank E Harrell Jr Professor and Chair School of Medicine Department of Biostatistics Vanderbilt University