topic Re: Proc mixed level 2 variable in Statistical Procedures

Proc mixed level 2 variable

PaulaC — Sat, 24 Jun 2017 03:05:24 GMT

I have two separate clinical trial datasets that have the same variables and sites are numbered consecutively for each clinical trial. I have combined both clinical trials into one dataset using the set statement. I also have a flag for each clinical trial. I have a unique subject id that is unique across both clincial trial datasets. I am clustering patients within sites. The sites ids are the same in many cases across both clinical trials. Is that a problem with proc mixed and using siteid as the clustering variable? Does siteid need to be unique in the combined dataset eventhough the patient id is unique? I have a feeling it is, but wanted to confirm.

I used the following code:

proc mixed data=analysisaim1 covtest method=ML order=data;
class siteid;
model chgql=trial racecat cntrycat totcomorbid suppcare ecogb NumPt trtdur/solution ddfm=kr;
random intercept/subject=siteid;
run;

I am using SAS 9.4 TS. Any help you could provide would be greatly appreciated. Thanks.

Re: Proc mixed level 2 variable

lvm — Sat, 24 Jun 2017 22:04:48 GMT

The side IDs should be unique for the nesting you describe. Easy to handle. First, trial should be listed in the class statement (I can't imagine how trial is a continuous variable). then in the random statement, use subject=siteid(trial);

Re: Proc mixed level 2 variable

PaulaC — Sat, 24 Jun 2017 23:44:55 GMT

Thank you for your help with this. I had copied and pasted from my code but did not copy the class statement properly. Trial is a categorical variable and not a continuous variable. I do have trial in my class statement. That is a pretty easy fix. I will try that. Thank you for your help.