Due to missing data with some of my predictor variables I first do multiple
imputation as follows:
library(foreign)
library(Amelia)
library(norm)
set.seed(666)
M=10
impdat <-
NVP[,c("X_t0","X_t","nvp","adstatus","t0rwfa","ageatran","whostage","t0rhfa","vlsupp","t0rwfh","t0rvl","t0rcd4pc","postrantb","resistance","id")]
round(apply(apply(impdat,c(1,2),is.na),2,mean),digits=3) # percentage of
missing values
myimp <- amelia(impdat, m=M, p2s=0,
noms=c("postrantb","vlsupp","whostage","resistance"),cs=c("id"),ts=c("X_t0"),bounds=matrix(c(3,0,70,
4,0,100, 11,0,400, 12,0,50
),ncol=3,nrow=4,byrow=T),logs="nvp",polytime=3,splinetime=3,empri=1000,incheck=TRUE,tolerance=0.001)
The I do cox model as follows
# Categorization of relevant variables
for(i in 1:10){
myimp$imputations[[i]] <-
cbind(myimp$imputations[[i]],cut(myimp$imputations[[i]]$t0rvl,breaks=c(-1,50,400,1000000000)))
colnames(myimp$imputations[[i]])[15]<-c("vlcat")
myimp$imputations[[i]] <-
cbind(myimp$imputations[[i]],cut(myimp$imputations[[i]]$t0rcd4pc,breaks=c(-1,25,1000000000)))
colnames(myimp$imputations[[i]])[16]<-c("cd4pccat")
myimp$imputations[[i]] <-
cbind(myimp$imputations[[i]],cut(myimp$imputations[[i]]$ageatran,breaks=c(-1,36,1000000000)))
colnames(myimp$imputations[[i]])[17]<-c("agecat")
myimp$imputations[[i]] <-
cbind(myimp$imputations[[i]],cut(myimp$imputations[[i]]$t0rwfa,breaks=c(-1000,-3,-2,10)))
colnames(myimp$imputations[[i]])[18]<-c("wfacat")
myimp$imputations[[i]]
<-cbind(myimp$imputations[[i]],cut(myimp$imputations[[i]]$t0rwfh,breaks=c(-1000,-3,-2,10))
colnames(myimp$imputations[[i]])[19]<-c("wfhcat")
myimp$imputations[[i]] <-
cbind(myimp$imputations[[i]],cut(myimp$imputations[[i]]$t0rhfa,breaks=c(-1000,-3,-2,10)))
colnames(myimp$imputations[[i]])[20]<-c("hfacat")
myimp$imputations[[i]] <-
cbind(myimp$imputations[[i]],cut(myimp$imputations[[i]]$nvp,breaks=c(-1,1,1000000000)))
colnames(myimp$imputations[[i]])[21]<-c("nvpcat1")
myimp$imputations[[i]] <-
cbind(myimp$imputations[[i]],cut(myimp$imputations[[i]]$nvp,breaks=c(-1,4,1000000000)))
colnames(myimp$imputations[[i]])[22]<-c("nvpcat10")}
head(myimp$imputations[[1]])
######
Then I perform cox regression as follows
m2_1<-coxph(Surv(X_t0,X_t, vlsupp) ~ nvp + as.factor(cd4pccat) +
as.factor(vlcat) + as.factor(agecat) + as.factor(whostage) +
as.factor(hfacat) + as.factor(wfacat) + as.factor(wfhcat) +
as.factor(resistance) + as.factor(postrantb) +
cluster(id),data=myimp$imputations[[1]],method="breslow",robust=TRUE)
summary(m2_1)
The I get the following eWarning message:
In fitter(X, Y, strats, offset, init, control, weights = weights, :
Ran out of iterations and did not converge
I have tried to increase the maximum number of iterations,however I still
get the same warning message.
I have also looked at change my initial estimates but I get the following
error
Error in fitter(X, Y, strats, offset, init, control, weights = weights, :
Wrong length for inital values
Does anybody know have any suggestions as to I can get the model to
converge
Kind Regards
--
Retsilisitsoe Moholisa, MSc Medical Biochemistry
PhD Student in Pharmacometrics
Pharmacometrics Lab
Department of Clinical Pharmacology
Faculty of Health Sciences
University of Cape Town
Tel:021-404 7719
Cell:0712801254
[[alternative HTML version deleted]]
______________________________________________
R-help@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.
