How to only exclude ID's with subsequent unwanted observations

[tomas-andriotti]

Hello, 

i need help with coding a longitudinal dataset.

I want to exclude individuals from a long format dataset which have different levels of the covariate over time, i.e., if they present a hospital type observations in one occasion as PFF system and in one another as DIRECT system, they should be excluded. if on all occasions they are only PFF or only DIRECT they should stay in the dataset. Thus, I can only keep those who do not cross between systems over time. The covariate is binary, PFF/DIREC.

Look forward to hearing from you.

TA

[PGStats · Accepted solution]

Try something like:

proc sql;
create table want as
select * 
from have
group by patientID
having count(distinct hospitalType) <= 1;
quit;

[tomas-andriotti]

thank you ! very helpful