using a macro to flag access observations outside the time period

[Bluebonnet16]

Hi,

I have a dataset with clients who had certain procedure performed over the 3 year time period. I need to flag those instances where  4 or more procedures took place outside the 180 day period for each client ID (status=flagged). Here is a sample data set.

ID	serv_dt	lag	PROCCD	proccdcnt	status
217207401	3/22/2017	0	95951	4
217207401	4/2/2018	376	95951	4
217207401	4/3/2018	1	95951	4
217207401	5/23/2018	50	95951	4
220135502	11/4/2016	0	95951	4
220135502	11/13/2016	9	95951	4
220135502	11/14/2016	1	95951	4
220135502	11/15/2016	1	95951	4	flagged
237287806	11/17/2017	0	95951	6
237287806	11/18/2017	1	95951	6
237287806	11/19/2017	1	95951	6
237287806	10/28/2018	343	95951	6
237287806	10/29/2018	1	95951	6
237287806	10/30/2018	1	95951	6
220462501	10/17/2018	0	95951	5
220462501	10/18/2018	1	95951	5
220462501	10/19/2018	1	95951	5
220462501	10/20/2018	1	95951	5
220462501	10/21/2018	1	95951	5	flagged
249892003	12/7/2016	0	95951	4
249892003	12/8/2016	1	95951	4
249892003	12/9/2016	1	95951	4
249892003	12/10/2016	1	95951	4	flagged
257527503	4/14/2017	0	95951	6
257527503	4/15/2017	1	95951	6
257527503	5/28/2017	43	95951	6
257527503	7/18/2018	416	95951	6
257527503	7/19/2018	1	95951	6
257527503	7/20/2018	1	95951	6
219482501	8/30/2016	0	95951	11
219482501	8/31/2016	1	95951	11
219482501	9/1/2016	1	95951	11
219482501	9/2/2016	1	95951	11	flagged
219482501	9/3/2016	1	95951	11	flagged
219482501	9/4/2016	1	95951	11	flagged
219482501	9/5/2016	1	95951	11	flagged
219482501	9/6/2016	1	95951	11	flagged
219482501	9/7/2016	1	95951	11	flagged
219482501	9/8/2016	1	95951	11	flagged
219482501	9/9/2016	1	95951	11	flagged

So, for example, the first ID=217207401 will not be flagged. But the second, id=220135502 will be flagged as there were 4 procedures done in less than 180 days.

I understand that there should probably be some way of using a macro to count within each ID group the number of days for the consequitive 4 observations, and flag them when it's more than 180. But I'm not sure how to do that as teh number of rows for each ID is different. Any help or advice would be appreciated.

Thank you!

[PaigeMiller · Accepted solution]

Macros and macro variables are most definitely not needed, and not worth the complication.

Just sort by ID and serv_dt and then (assuming serv_dt is really a SAS date and not a character string).

data want;
    set have;
    prev_id=lag3(id);
    prev_dt=lag3(serv_dt);
    if prev_id=id and serv_dt-prev_dt<180 then status='flagged';
run;

[Bluebonnet16]

Thank you so much!! It worked great and I understand the logic behind it. Thank you!!