topic Re: Subsequent date-interval computation in SAS Programming

Subsequent date-interval computation

cj3 — Tue, 12 Apr 2022 16:09:52 GMT

Hello Community,

I need to create a variable in my dataset to indicate whether or not a patient received a second prescription of a medication within 34 days of their initial prescription (i.e., the patient has one or more unique prescription dates within 34 days of the index prescription date). Please see below for how my dataset is currently organized and how I would like the output to be like.

Note that my data is already sorted ascending by prescription date, therefore the first date listed for each patient is the index date (i.e., the first prescription received). Also note that sometimes multiple rows of data are entered for the same prescription, which can result in the same prescription date shown more than once for a patient. Therefore, it is important that the code identifies whether a UNIQUE date is within 34 days of the index date. Finally, please note that patients who did not receive the medication will have the prescription_date variable equal to missing. In these cases, it should be new_var=0 since the condition would be false for these patients as well. Any help with the coding for doing this would be greatly appreciated!!

Have:

Patient_id	Prescription_date
001	17JUN2020
001	23JUL2020
001	31JUL2020
002	15FEB2020
002	15FEB2020
002	30MAR2020
003	02JAN2021
003	02JAN2021
004	21APR2021
004	20JUN2021
004	01JUL2021
005	19OCT2021
005	25OCT2021
005	25OCT2021
005	28SEP2021
006	.

Want:

Patient_id	New_var
001	1
002	0
003	0
004	0
005	1
006	0

Re: Subsequent date-interval computation

Kurt_Bremser — Tue, 12 Apr 2022 16:22:38 GMT

See this:

data have;
input Patient_id $;
input Prescription_date :date9.;
format Prescription_date yymmdd10.;
datalines;
001
17JUN2020
001
23JUL2020
001
31JUL2020
002
15FEB2020
002
15FEB2020
002
30MAR2020
003
02JAN2021
003
02JAN2021
004
21APR2021
004
20JUN2021
004
01JUL2021
005
19OCT2021
005
25OCT2021
005
25OCT2021
005
28SEP2021
006
.
;

data want;
set have;
by patient_id;
retain new_var index;
if first.patient_id
then do;
  index = prescription_date;
  new_var = 0;
end;
else if 1 le prescription_date - index le 43 then new_var = 1;
if last.patient_id;
keep patient_id new_var;
run;

Supplying example data in a data step with datalines makes helping you easier for us.

Re: Subsequent date-interval computation

cj3 — Wed, 13 Apr 2022 13:08:16 GMT

@Kurt_Bremser This is perfect. Thank you!!