I have a set of patients who have hospital visits at different ages. Each patient may have hospital visit more than once. I want to flag THOSE INDIVIDUALS(IDs) who have ONLY hospital visits at the age of less than 18. If an individual has hospital visits at age below 18 and above 18, I don’t want to flag them.
For example:
Data I have:
Patient_ID | Age |
1 | 14 |
1 | 16 |
1 | 18 |
2 | 15 |
2 | 17 |
3 | 16 |
4 | 19 |
Data I Want
Patient_ID | Age | Flag |
1 | 14 | 0 |
1 | 16 | 0 |
1 | 18 | 0 |
2 | 15 | 1 |
2 | 17 | 1 |
3 | 16 | 1 |
4 | 19 | 0 |
Thank you very much.
Hi @SR11 Since your dataset seems to be nicely sorted by Patient_id age in ascending order, makes it a straight forward boolean-
data have;
input
Patient_ID
Age ;
cards;
1
14
1
16
1
18
2
15
2
17
3
16
4
19
;
data Want;
do _n_=1 by 1 until(last.patient_id);
set have;
by patient_id;
flag=age<18;
end;
do _n_=1 to _n_;
set have;
output;
end;
run;
proc sql;
create table want as
select * , flag = max(age) < 18
from have
group by patient_id
order by patient_ID, age;
quit;
The logic simplifies to the maximum age for a person must be less than 18. You can use SQL to find the maximum age per ID (group by) and then use it to calculate your flag.
@SR11 wrote:
I have a set of patients who have hospital visits at different ages. Each patient may have hospital visit more than once. I want to flag THOSE INDIVIDUALS(IDs) who have ONLY hospital visits at the age of less than 18. If an individual has hospital visits at age below 18 and above 18, I don’t want to flag them.
For example:
Data I have:
Patient_ID
Age
1
14
1
16
1
18
2
15
2
17
3
16
4
19
Data I Want
Patient_ID
Age
Flag
1
14
0
1
16
0
1
18
0
2
15
1
2
17
1
3
16
1
4
19
0
Thank you very much.
Hi @SR11 Since your dataset seems to be nicely sorted by Patient_id age in ascending order, makes it a straight forward boolean-
data have;
input
Patient_ID
Age ;
cards;
1
14
1
16
1
18
2
15
2
17
3
16
4
19
;
data Want;
do _n_=1 by 1 until(last.patient_id);
set have;
by patient_id;
flag=age<18;
end;
do _n_=1 to _n_;
set have;
output;
end;
run;
And just in case if your age variable isn't sorted-
data Want;
do _n_=1 by 1 until(last.patient_id);
set have;
by patient_id;
flag=flag<>age;
end;
flag=flag<18;
do _n_=1 to _n_;
set have;
output;
end;
run;
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.