Joint model

[meleseenyew1919]

Hello @sbxkoenk! Have you seen the article! And so far I can't get any solution.
I can't access this file/ your first solution:
https://communities.sas.com/t5/Statistical-Procedures/Joint-model/m-p/741898#M36084

[sbxkoenk]

Hello,

How do you mean: "You can't open / access it" ??

You are in it.

You are on the second page. Scroll to the bottom and click on "1" (page 1). There's a button bar showing a "1" and a "2" (page 1 and 2, respectively).

Click on 1 and you will see all the earlier posts.

Koen

[meleseenyew1919]

I can't open this, is there any attachment on View solution in original post?

[sbxkoenk]

Hello @meleseenyew1919 ,

I contacted a colleague meanwhile.

She says:

=========== 

GLIMMIX can estimate ordinal model using DIST=MULTINOMIAL LINK=CLOGIT. 
However, GLIMMIX only supports G-side random effects for Multinomial models. In other words, you can not model R-side random effects (correlated errors for repeated measures) for Multinomial models in PROC GLIMMIX.

I do not have any examples of joint modeling of longitudinal and time-to-event (survival) data.

=========== 

[Later edit]

Also have a look at this entry:

Looking for SAS code for Joint Model between survival and longitudinal analysis model
https://communities.sas.com/t5/SAS-Health-and-Life-Sciences/Looking-for-SAS-code-for-Joint-Model-bet...

Koen

[meleseenyew1919]

Hello Koen! sorry for that! on google, the question is put as a solved
question I know that is my mistake ( marked my first "answer" as a
solution, because this is my first question in SAS) so I would like to say
I can't rich the answer!

PROC NLMIXED is best for ordinal rather than PROC GLIMMIX as some research
states. so if you can give me the last chance with a great excuse I don't
get the code so far.



Thank you so much for your patience and all helps🙏!

[sbxkoenk]

Hello @meleseenyew1919 ,

No need to apologize obviously. 😉

Concerning PROC NLMIXED: I have never used PROC NLMIXED before and have no time to look into it this week.

I suggest you start a new topic (a new question) in this same board (Home > Analytics > Statistical Procedures).

That way it becomes a fresh / new topic that will be looked at by almost "everyone" following the board, which maximizes your chances for success. As told before, this topic is no more looked at because it seems solved for the others.

Good luck,

Koen