I am using the National Health Interview Survey to look at cardiovascular disease rates by various demographic groups. I have multiply imputed my data by FCS method through PROC MI without any issue. Next, I would like to calculate age-adjusted prevalence rates for cardiovascular disease by various demographic groups. I have been using PROC SURVEYREG with ESTIMATE statements. When I use PROC MIANALYZE, I obtain an error message that reads, "ERROR: The model effect estimate is not in the PARMS= data set" even though it clearly is. I have added my SAS code below.
**MULTIPLE IMPUTATION**;
*Check how much data is missing*;
proc mi data=nhis nimpute=0 ;
var CVD AGEP_A Sex RaceEthnicity Education IncomePovertyRatio BMICategory HealthInsurance
URBRRL REGION;
ods select misspattern;
run;
*Imputation phase*;
proc mi data=nhis nimpute=5 out=mi_fcs ;
class CVD Sex RaceEthnicity Education IncomePovertyRatio BMICategory HealthInsurance URBRRL REGION;
fcs plots=trace(mean std);
var CVD AGEP_A Sex RaceEthnicity Education IncomePovertyRatio BMICategory HealthInsurance URBRRL REGION;
fcs discrim(CVD Sex RaceEthnicity Education IncomePovertyRatio BMICategory HealthInsurance URBRRL REGION /classeffects=include) nbiter =10 ;
run;
**Analysis phase for age-adjusted prevalence**;
proc surveyreg data=work.mi_fcs;
strata PSTRAT;
cluster PPSU;
weight WTFA_A;
class AgeCat;
domain _imputation_*RaceEthnicity;
model CVD2 = AgeCat / noint solution clparm;
estimate 'Total ' AgeCat 0.2015 0.1733 0.1568 0.1566 0.1551 0.1566;
ods output ParameterEstimates=estimates (where=(_imputation_ ne . ));
run;
proc mianalyze parms(classvar=classval)=estimates;
by RaceEthnicity;
modeleffects estimate ;
stderr StdErr;
run;
**In this last step with proc mianalyze, I encounter the error message that the model effect estimate is not in the parms= data set**
Would you please offer any support or guidance to resolve this issue? Thank you!