Hi,
I have claims dataset that I need to calculate duration of continuous medication use for 90 days.
My dataset looks like this:
Pat_ID fill_date Medication day supply
1 01/01/2017 A 15
1 02/01/2017 B 20
1 03/01/2017 A 30
1 04/01/2017 A 30
1 05/01/2017 A 30
I am not sure if it is possible to achieve a result looking like this
Pat Id Medication day supply continuous day use
1 A 15 15
1 B 20 20
1 A 30 90
1 A 30 90
1 A 30 90
Thanks much
Not directly with the info provided. I would first assign a intake number, then you can simply group and sum and merge back on, something like (and not tested as i don't type test data in):
data have; set have; by pat_id; if first.pat_id then do; start_date=fill_date; end_date=fill_date; continuous=day_supply; end; else if medication ne lag(medication) or fill_date ne lag(fill_date)+1 then do; output; start_date=fill_date; end_date=fill_date; continuous=day_supply; end; else do; end_date=fill_date; continuous=sum(continuous,day_supply); end; run;
Basically keep a count of start_date and end_date and sum(day_intake) for each group, and output one record at the end of the group - note you need start and end to be able to merge sum() back to the data.
It's finally time to hack! Remember to visit the SAS Hacker's Hub regularly for news and updates.
Learn how use the CAT functions in SAS to join values from multiple variables into a single value.
Find more tutorials on the SAS Users YouTube channel.
Ready to level-up your skills? Choose your own adventure.