Selecting observations

[JKUW]

Study ID	Drug1	Drug2	Drug3
00001	1	0	0
00001	0	0	1

Hello,

I just started using SAS again and I have yet to refresh my memory.

I'm currently investigating 3 drugs - patients are only supposed to be prescribed to only 1 drug.

However, some patients received more than 1 drug (as depicted in the table above) and I want to delete these specific individuals from the data.

Any codes to identify and delete these individuals? Any advice/help would be appreciated.

[novinosrin]

Hi, There's studyid but where is Patient/individual id? Please post a more comprehensive sample of what you have and want. Thank you

[JKUW]

Hi novinosrin,

Study_ID in this case would be the patient_id.

[unison]

Maybe this:

data have;
	format studyid z5.;
	input studyid drug1-drug3;
	datalines;
00001 1 0 0
00001 0 0 1
00002 0 1 0
;
run;

data want;
	set have;
	by studyid;

	if first.studyid then
		rx_total=0;
	rx_total+sum(of drug:);

	if last.studyid and rx_total=1 then
		output;
	drop rx_total;
run;

assumes only one drug can be 1 for any given observation.