Hi Rob,   Thanks for this information. This was how I had run my other models, so it's good to know I did it correctly. I was trying to utilize the covariance matrix and parminfo table to create confidence intervals for odds ratios involving my interaction term. It doesn't seem like there is code to do this within proc mianalyze (such as estimate or contrast statements) and the covariance table is by imputation. I am starting to think this might be too complex for me, and my best option might be to simply run stratified models, rather than using an interaction term. Unless there is some way to combine the covariances from the various imputations, similar to the way proc mianalyze combines the estimates from the imputations?   If you have any further thoughts on this I would welcome them, otherwise thanks for your help.   Katy ... View more

Proc gee data = /*mizheight descending*/ diss.aim1_mi_analysis descending /*small_mi 13 ! descending*/ ; NOTE: Writing HTML Body file: sashtml.htm 14 where age_mo <= &age; 15 weight weight; 16 by _imputation_; 17 Class psnum visit haz4 (ref = '3') season (ref = first) sector (ref = first) 18 hbasch (ref = 'AA') /param = ref; 19 Model &malvar = haz4 nmanem4 laginfection season momed ses2 age_mo age_mosq age_mocu 20 malaria sector mompar hbasch sexa lagzheight lag2zheight mal1prev mal2prev 20 ! haz4*nmanem4 / dist = mult link = glogit; 21 Repeated subject = psnum / within = visit ECOVB; 22 ods output GEEEmpPEst=parm_zheight_int_nma_&malvar ParmInfo= parminfo GEERCov= 22 ! cov_&malvar; 23 run; NOTE: The data set WORK.COV_MAL1 has 2800 observations and 58 variables. NOTE: The data set WORK.PARM_ZHEIGHT_INT_NMA_MAL1 has 2800 observations and 10 variables. WARNING: Output 'ParmInfo' was not created. Make sure that the output object name, label, or path is spelled correctly. Also, verify that the appropriate procedure options are used to produce the requested output object. For example, verify that the NOPRINT option is not used. NOTE: PROCEDURE GEE used (Total process time): real time 59.64 seconds cpu time 41.01 seconds ... View more

Hi Reeza,   Yes, I see the covariance matrix for each imputation in the output.   Katy ... View more