topic Re: BINARY REPEATED MEASURES IN GLIMMIX FOR ADJUST P-VALUE in SAS Programming

BINARY REPEATED MEASURES IN GLIMMIX FOR ADJUST P-VALUE

superbibi — Fri, 07 Jun 2019 19:08:48 GMT

Hi SAS users,

I have a data set with binary repeated measure outcome.

I want to know whether the incidence of GSRSINCI6 does not change over visits. Also, I want to know the difference (p-value) of each visit compared to visit 2.

My question is whether there is a way that I can adjust the p-value for multiple between visit comparison in proc glimmix.

My code and data set have been uploaded.

Thanks.

PROC GLIMMIX DATA = GSRS;
CLASS SUBNUM VISIT (REF = '2') ;
MODEL GSRSINCI6(EVENT ='1') = VISIT/DIST= BINARY SOLUTION CHISQ ODDSRATIO ;
RANDOM _RESIDUAL_ /SUBJECT = SUBNUM TYPE = VC ;
RUN;

Re: BINARY REPEATED MEASURES IN GLIMMIX FOR ADJUST P-VALUE

PGStats — Sat, 08 Jun 2019 04:01:39 GMT

I guess you may want:

PROC GLIMMIX DATA = sasforum.bin;
CLASS ID VISIT (REF = '2') ;
MODEL GSRSINCI6(EVENT ='1') = VISIT/DIST= BINARY SOLUTION CHISQ ODDSRATIO ;
RANDOM _RESIDUAL_ /SUBJECT = ID TYPE = VC ;
lsmeans visit / diff=control('2') adjust=dunnett;
RUN;

Re: BINARY REPEATED MEASURES IN GLIMMIX FOR ADJUST P-VALUE

superbibi — Wed, 12 Jun 2019 14:17:43 GMT

Thank you. But since the data is binary, is it appropriate to use lsmeans?

Re: BINARY REPEATED MEASURES IN GLIMMIX FOR ADJUST P-VALUE

PGStats — Wed, 12 Jun 2019 21:41:13 GMT

Oops, I'm sorry. I guess not.

You can always adjust multiple p-values (or a subset of them) with proc multtest after exporting with ods ouitput.