Hi @derbygun,   To elaborate on the point about correlation from @StatDave : do you have multiple (repeated) measurements on each person_id? If not, there's nothing correlated to model, so we'd omit person_id from the model.   Since observations within zip codes are presumably correlated, you could use model #1.   More details on Proc Genmod Repeated parameterization here. Reading this from your example, person_id would be equivalent to patient and zipcode would equate to physician.   I favor Proc GLIMMIX because it's intuitive to input natural hierarchies, for example:   random person_id / subject=zipcode residual ; and it uses likelihood techniques which some practitioners hold to be preferable to moment specifications like those in Genmod.   Good luck! ... View more

data have; input ( PersonID VisitID NamesofMedications Startdate DrugclassA DrugclassB DrugclassC ) (: $20.); cards; 2 3866897 dula 28-Nov-18 0 1 0 2 6139545 dula 19-Jun-19 0 1 0 2 14110036 dula 7-Oct-20 0 1 0 4 3866996 dapa-met 12-Nov-18 1 0 0 4 3866996 Insu-lixi 12-Nov-18 0 1 0 4 6048002 dula 8-Jun-20 0 1 0 4 60410643 dula 5-Oct-20 0 1 0 6 1453918 Sita 22-Mar-17 0 0 1 6 2470214 dula 21-May-18 0 1 0 6 3866906 dula 17-Dec-18 0 1 0 ; proc sql; create table want as select * from have where PersonId not in (select distinct PersonID from have group by PersonID,StartDate having count(distinct NamesofMedications) ne 1); quit; ... View more