R help - Aug 2011 - Repeated measures cummulative logit mixed model

Dear R help gurus, 


I have the following problem and I would be delighted if you could help me.
>From a large (1500) cohort of patients we have been taking some measurements
(ECG measurements, but its not important). The measurements are ordinal in 4
grades (Grade I-IV, grade IV being the most severe form). Every patients has
been measured several times (usually once per year). The follow-ups are
different for each patient (ranging from days to several years), thus some
patients have only 1 measurement and some have several. A patient can shift
from Grade to Grade, that is a patient can be measured now as Grade I and
later as Grade II and then at then next measurement he can return to Grade
I.

My data are in the following format

Patient ID, Time of Measurement, Grade

I would like to model the probability of being in each of the Grades (I-V)
as a function of time. I guess i require a repeated measure cumulative logit
mixed model with random slope and intercept (for each patient).

Could anyone pin point me at any information that would help me accomplish
this in R? Packages? Books?

Information about categorical data analysis with R seems to be poor in the
Internet.

Thank you very much on advance
JT

--
View this message in context:
http://r.789695.n4.nabble.com/Repeated-measures-cummulative-logit-mixed-model-tp3746845p3746845.html
Sent from the R help mailing list archive at Nabble.com.

On Aug 16, 2011, at 6:36 AM, Jim Trabas wrote:
> Dear R help gurus,
>
>
> I have the following problem and I would be delighted if you could  
> help me.
>
>> From a large (1500) cohort of patients we have been taking some  
>> measurements
> (ECG measurements, but its not important). The measurements are  
> ordinal in 4
> grades (Grade I-IV, grade IV being the most severe form). Every  
> patients has
> been measured several times (usually once per year). The follow-ups  
> are
> different for each patient (ranging from days to several years),  
> thus some
> patients have only 1 measurement and some have several. A patient  
> can shift
> from Grade to Grade, that is a patient can be measured now as Grade  
> I and
> later as Grade II and then at then next measurement he can return to  
> Grade
> I.
This should probably be followed up on the mixed models SIG list.

(Using Baron's RSiteSearch page.)

Search results with : repeated "proportional odds logistic"

http://finzi.psych.upenn.edu/R/library/repolr/html/00Index.html
http://finzi.psych.upenn.edu/R/Rhelp02/archive/99576.html

Search results with: repeated "ordered logit"

http://finzi.psych.upenn.edu/R/library/ordinal/doc/primer.pdf

That last one has citations to Agresti articles, which means it may  
have useful termininogy to be used in searching the very informative  
Thompson/Agresti compendium of R methods:

https://home.comcast.net/~lthompson221/Splusdiscrete2.pdf

The mixed-models SIG would seem to be a natural place to look and for  
that I use the advanced search panel fpr google to construct this  
search of its archives:

http://www.google.com/search?hl=en&lr=&q=repeated+logit+site%3Ahttps%3A%2F%2Fstat.ethz.ch%2Fpipermail%2Fr-sig-mixed-models

(I did trim some of the trailing bits that seemsed superfluous.)

-- 
David.
> My data are in the following format
>
> Patient ID, Time of Measurement, Grade
>
> I would like to model the probability of being in each of the Grades  
> (I-V)
> as a function of time. I guess i require a repeated measure  
> cumulative logit
> mixed model with random slope and intercept (for each patient).
>
> Could anyone pin point me at any information that would help me  
> accomplish
> this in R? Packages? Books?
>
> Information about categorical data analysis with R seems to be poor  
> in the
> Internet.
>
> Thank you very much on advance
> JT
>
> --
> View this message in context:
http://r.789695.n4.nabble.com/Repeated-measures-cummulative-logit-mixed-model-tp3746845p3746845.html
> Sent from the R help mailing list archive at Nabble.com.
>
> ______________________________________________
> 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.
David Winsemius, MD
West Hartford, CT