Simple macro but doesn't work

[Cruise]

I'm trying to repeat the code below 20 times for 20 consequent years. The macro doesn't help. The problem I think is

in the "date_of_diagnosis_yyyy='n' " . 

Please help in the code, if you have time. Any idea why this approach doesn't work?

%macro models(param,par,n);

ods output parameterestimates=&param;
proc genmod data=have;
class zone(ref="0") sex(ref="1") racecat(ref="1") agecat(ref="2")/param=ref;
model n_cases_reg=zone sex racecat agecat
      zone|date_of_diagnosis_yyyy/ dist=negbin link=log offset=ln type3;
ods trace on;
where zone ne '9' and date_of_diagnosis_yyyy eq 'n';
run;  

ods rtf close;
data ∥ set &param;
if pam~ in ("Scale","Intercept") then do;
RR= exp(estimate);
low_rr=exp(estimate-(1.96*stderr));
up_rr=exp(estimate+(1.96*stderr));
end;
run;

%mend models;
%models(param1995,par1995,1995);
.
so forth so on until
.
%models(param2015,par2015,2015);

[Astounding]

The problem does not lie within the macro.  The problem lies with the fact that the macro exists at all.  If you want a separate analysis for each year, sort your data by year and run GENMOD by year.

[Cruise]

I don't quite get it Astounding.

[Reeza]

Your parameters are macro variables. Remember they only resolve in double quotes and need an ampersand. The line you indicated does neither.

[Reeza]

 date_of_diagnosis_yyyy eq “&n”

[Reeza · Accepted solution]

ODs output ....; proc genmod data=have; By date_of_diagnosis_yyyy; class zone(ref="0") sex(ref="1") racecat(ref="1") agecat(ref="2")/param=ref; model n_cases_reg=zone sex racecat agecat zone|date_of_diagnosis_yyyy/ dist=negbin link=log offset=ln type3; ods trace on; where zone ne '9'; run; 

Any reason you can’t run this as above?