Identifying overlap medication use

[riya275]

I have a problem similar to the problem in -

https://communities.sas.com/t5/SAS-Programming/Concomitant-drug-medication-use/m-p/339879#M77587

However I have an issue , I have overlapping of same drug as well -

Eg:

ID	DRUG	START_DT	DAYS_SUPP	END_DT
1	A	2/17/2010	30	3/19/2010
1	A	3/17/2010	30	4/16/2010
1	A	4/12/2010	30	5/12/2010
1	A	8/20/2010	30	9/19/2010
1	B	5/6/2009	30	6/5/2009

Here the three A prescriptions are over lapping . 

So using the code in the link gives me combinations like A-A-B

whereas I dont want that.

However I want to account for the overlapping days for drug A. So I want to shift the second row prescription to 3/20/2010 to 4/19/2010. Similarly for 3rd A prescription. 

[PaigeMiller]

---

@riya275 wrote:

I have a problem similar to the problem in -

https://communities.sas.com/t5/SAS-Programming/Concomitant-drug-medication-use/m-p/339879#M77587

However I have an issue , I have overlapping of same drug as well -

Eg:

ID	DRUG	START_DT	DAYS_SUPP	END_DT
1	A	2/17/2010	30	3/19/2010
1	A	3/17/2010	30	4/16/2010
1	A	4/12/2010	30	5/12/2010
1	A	8/20/2010	30	9/19/2010
1	B	5/6/2009	30	6/5/2009

Here the three A prescriptions are over lapping . 

So using the code in the link gives me combinations like A-A-B

whereas I dont want that.

However I want to account for the overlapping days for drug A. So I want to shift the second row prescription to 3/20/2010 to 4/19/2010. Similarly for 3rd A prescription. 

---

Try this:

data want;
    set have;
    prev_end_dt=lag(end_dt);
    if start_dt<=prev_end_dt then start_dt=prev_end_dt+1;
    drop prev_end_dt;
run;

This assumes that START_DT and END_DT are actual SAS date values, formatted as shown.

[riya275]

 does not take into account the days of supply. His total days of supply will reduce this way

[PaigeMiller]

---

@riya275 wrote:

 does not take into account the days of supply. His total days of supply will reduce this way

---

Please give an example.

[ed_sas_member]

Hi @riya275 

Could you please test if the following approach meet your needs?

data have;
	infile datalines dlm="09"x;
	input ID DRUG $ START_DT: MMDDYY10. DAYS_SUPP END_DT:MMDDYY10.;
	format START_DT END_DT MMDDYY10.;
	datalines;
1	A	2/17/2010	30	3/19/2010
1	A	3/17/2010	30	4/16/2010
1	A	4/12/2010	30	5/12/2010
1	A	8/20/2010	30	9/19/2010
1	B	5/6/2009	30	6/5/2009
;
run;

proc sort data=have out=have_sorted;
	by ID DRUG START_DT;
run;

data have2;
	set have_sorted;
	format NEW_START_DT NEW_END_DT _lagEND_DT MMDDYY10.;
	
	_lagID = lag(ID);
	_lagDRUG = lag(DRUG);
	_lagEND_DT = lag(END_DT);
	
	if ID = _lagID and DRUG = _lagDRUG and START_DT <= _lagEND_DT then flag=1;
	else flag = 0;
	
	retain NEW_START_DT NEW_END_DT;

	if flag=0 then do;
		NEW_START_DT = START_DT;
		NEW_END_DT = END_DT;
	end;
	else do;
		NEW_START_DT + DAYS_SUPP + 1;
		NEW_END_DT + DAYS_SUPP + 1;
	end;
	drop flag _:;
run;

 

[riya275]

This gives incorrect results if the days_supp is different for rows. So I have modified it a bit -

data have2;
set have_sorted1;
format NEW_START_DT NEW_END_DT _lagEND_DT date9.;

_lagID = lag(patient_ID);
_lagDRUG = lag(drg_cls);
_lagEND_DT = lag(rx_ed_dt);

if patient_ID = _lagID and drg_cls= _lagDRUG and rx_st_dt <= _lagEND_DT then flag=1;
else flag = 0;

retain NEW_START_DT NEW_END_DT;

if flag=0 then do;
NEW_START_DT = rx_st_dt;
NEW_END_DT = rx_ed_dt;
end;
else do;
New_start_dt = NEW_End_DT + 1;
NEW_END_DT = new_start_dt + DAY_SUPP ;
end;
/* drop flag _:;*/
run;

But even then I get incorrect result -

Patient_id	Drug	drug_start	day_supp	drug_end	New_start	New_end
15	A	6-Sep-15	30	5-Oct-15	6-Sep-15	5-Oct-15
15	A	24-Sep-15	90	22-Dec-15	6-Oct-15	4-Jan-16
15	A	6-Dec-15	90	4-Mar-16	5-Jan-16	4-Apr-16
15	A	26-Feb-16	90	25-May-16	5-Apr-16	4-Jul-16
15	A	29-May-16	90	26-Aug-16	29-May-16	26-Aug-16
15	A	7-Dec-16	90	6-Mar-17	7-Dec-16	6-Mar-17
15	A	17-Feb-17	90	17-May-17	7-Mar-17	5-Jun-17

[riya275]

Found the solution-

https://stackoverflow.com/questions/60448872/identifying-overlap-medication-use