I'v downloaded STATS COMPRISK extension. the first time i run it i got the message "The first word in the line is not recognized as an SPSS Statistics command". I tried to delete the first line or add BEGIN PROGRAM R. but it didnt worked ( the message the first line...disapeared). how could i operate the command?

That message means that STATS COMPRISK was not installed. Did you install it or just download? If you installed it, you should have seen a message in the Viewer about the outcome of the install.

In some situations, for security reasons, you need to start Statistics using Run as Administrator or equivalent on Mac for the install to be successful.

Thank you.

I tried to run the command and got:

"The R cmprsk package is required but could not be loaded."

I downloaded the cmprsk package for R and insalled it and the same for SPSS. What can I do?

Thank you

Well, that is progress, since STATS COMPRISK at least started executing.

My guess is that the cmprsk package might have been installed in a different version of R from the one Statistics is using on your system.

Try running this code after starting Statistics as Administrator

begin program r.

install.packages("cmprsk")

end program.

The cmprsk package depends on the survival package. That should be installed automatically along with cmprsk, but if it isn't (try library(survival)), you can use install.packages to add it.

Thanks.

I tried to run the code :"

begin program r.

install.packages("cmprsk")

end program."

and got:

>Error # 1. Command name: end

>The first word in the line is not recognized as an SPSS Statistics command.

>Execution of this command stops

What should I do?

Thank you!

I also notice that the 'R' in the first line does not highlight like the rest.

Thanks!

The r in begin program r. won't highlight like the command name as it is a parameter value.

However, the error message indicates that the program did not get into R mode. Be sure that you highlight everything from begin to end program before running it (or use Run > All).

If the R code actually runs, you will be prompted to pick a CRAN site from a list that pops up.

Thank you!

I tried it and picked a CRAN site. Whats next?

It should have then installed the cmprsk module, and you should be able to run your command. However, if it didn't also installed the survival module that cmprsk needs, you will have to run the install.packages program for that one.

I got:"

begin program r. 

install.packages("cmprsk") 

end program. 

Installing package into 'C:/ProgramData/IBM/SPSS/Statistics/23/extensions' 

(as 'lib' is unspecified) 

Warning: unable to access index for repository https://cloud.r-project.org/bin/windows/contrib/3.1 

Error in readRDS(file) : 

 cannot read workspace version 3 written by R 4.2.0; need R 3.5.0 or newer 

In addition: Warning message: 

package 'cmprsk' is not available (as a binary package for R version 3.1.3)

I am guessing from the "installing package" message above that you are using Statistics V23, which is quite old. That would mean using R3.1, which is consistent with the error message. It appears that you may have an R4.2 distribution being used somewhere in tis process. That won't work. If so, you will need to figure out how that distribution is being referenced with V23. It might be that you have too new a version of cmprisk or survival, but normally CRAN would provide the package version that goes with the R version that is running. Statistics 23 will not work with R4.x versions.

Hellow. Iv downloaded SPSS V25 and the correct R version. When i tried to run the code I got:

"Error in data.frame(df[noncatcovar], catdf) :

arguments imply differing number of rows: 1830, 1794"

What can I do?

Thank you

I cannot help with this without details of the syntax you ran. It would also help to have the data if you can share it (jkpeckSupport Member).

This is the syntax:

DATASET ACTIVATE DataSet1.

STATS COMPRISK FAILTIME=T_TOT_SPP FAILSTATUS=Cause_Death_GROUP FIXEDCOVAR=SEX norigin agephase2 

  ishunll MCA_GROUP 

FAILCODE=2 CENSORCODE=1

/OPTIONS MISSING=OMIT PLOT=YES QUANTILES=.25 .50 .75 MAXITER=10 TOL=.000001.

I have 1830 subjects

Thank you

I think this has to do with missing data. Try running with MISSING=FAIL to check this.

If that is correct. try selecting only complete cases and running the command like this.

COUNT missingcount=your variables(MISSING).

temporary.

select if missingcount eq 0.

stats comprisk ...

If this is the solution, it would help me to improve the procedure if I could get the data (jkpeckSupport Member).

Using data->select cases i selected only the valid cases and ignored the missing data and it worked!

Thank you so much!

Hi, how can I calculate the confidance intervals using the cmprsk package? there isnt any option to choose the presentation of the CI.

The plot describes the hazard function for the two compiting risk and the censored? How can i tell which one is my primary outcome?

Thank you

I have added the CI to the Coefficients table. You can get it from here

https://1drv.ms/u/s!AoWcE61g_FAdisQJNGzg3ezH_EsUyQ?e=2Xgqq4

and install it over the existing version using Extensions > Install local extension bundle.

This will eventually get posted on the Extension Hub, but I would like to fix the missing data problem first. It would really help if you could send me the data and syntax where that occurred so that I can reproduce the problem (jkpeckSupport Member)

And the plot lines are annotated in the covariate evaluation table above the plot.

Thank you so much for all of your help! unfortunately I cannot send the file due to hospital policies. I attach the syntax.

STATS COMPRISK FAILTIME=T_TOT_SPP FAILSTATUS=Cause_Death_GROUP FIXEDCOVAR=agephase2 SEX_NEW norigin 

  ishunll TTG DIab_Group_N MCA_GROUP BP_GROUP2 BMI_GROUP TCH_GROUP 

FAILCODE=2 CENSORCODE=1

/OPTIONS MISSING=OMIT PLOT=YES QUANTILES=.25 .50 .75 MAXITER=10 TOL=.000001.

STATS COMPRISK FAILTIME=T_TOT_SPP FAILSTATUS=Cause_Death_GROUP FIXEDCOVAR=agephase2 SEX_NEW norigin 

  ishunll TTG DIab_Group_N MCA_GROUP BP_GROUP2 BMI_GROUP TCH_GROUP 

FAILCODE=2 CENSORCODE=1 

/OPTIONS MISSING=OMIT PLOT=YES QUANTILES=.25 .50 .75 MAXITER=10 TOL=.000001. 

Error in data.frame(df[noncatcovar], catdf) : 

 arguments imply differing number of rows: 1830, 1795

The porblem was fixed after I ignored the missing values.

Thank you

That's okay. I was able to reproduce the problem and fix it. The updated version, which also allows the user to control the confidence interval probability, can be obtained here.

https://1drv.ms/u/s!AoWcE61g_FAdisQJNGzg3ezH_EsUyQ?e=QP6OTB