Deriving analysis flag variable

[smackerz1988]

I have an analysis flag variable specification

Set to "Y" for all records across a subject/parameter category (USUBJID/PARCAT1N) that has at least one record where A1IND is "High", or A2IND is "Change high".

Here is a sample of the data and highlighted in red the desired output

USUBJID	A1IND	PARAMCD	ADT	ANL03FL
1001	Within range	SYSBP	20Apr2022
1001	Within range	SYSBP	10May2022
1001	Within range	SYSBP	26May2022
1001	Within range	SYSBP	01Jun2022
1001	High	DIABP	20Apr2022	Y
1001	Within range	DIABP	10May2022	Y
1001	Within range	DIABP	26May2022	Y
1001	Within range	DIABP	01Jun2022	Y
1001	No criteria defined for parameter	HR	20Apr2022
1001	No criteria defined for parameter	HR	10May2022
1001	No criteria defined for parameter	HR	26May2022
1001	No criteria defined for parameter	HR	01Jun2022
1001	No criteria defined for parameter	RESP	20Apr2022
1001	No criteria defined for parameter	RESP	10May2022
1001	No criteria defined for parameter	RESP	26May2022
1001	No criteria defined for parameter	RESP	01Jun2022
1001	No criteria defined for parameter	TEMP	20Apr2022
1001	No criteria defined for parameter	TEMP	10May2022
1001	No criteria defined for parameter	TEMP	26May2022
1001	No criteria defined for parameter	TEMP	01Jun2022
1001	No criteria defined for parameter	WEIGHT	20Apr2022
1001	No criteria defined for parameter	WEIGHT	10May2022
1001	No criteria defined for parameter	WEIGHT	01Jun2022
1001	No criteria defined for parameter	HEIGHT	20Apr2022
1001	No criteria defined for parameter	BMI	20Apr2022

[PaigeMiller]

Since your words talk about A2IND but that is not in your data set, I only answer the part about A1IND.

proc sql;
    create table want as select 
        *
        ,sum(a1nd='High')>0 as ANL03FL
    from have
    group by usubjid;
quit;

Also, since I prefer flag variables to be numeric 0 or 1, rather than "Y" or missing, I have programmed it to give you numeric 0s or 1s.

In the future, please do not provide data as screen captures, we cannot work with screen captures. In the future, please provide data as working SAS data step code, every single time, without us having to ask. Thank you.

[smackerz1988]

Thanks for this but this but I need it to only populate for all records of the parameter that contains one record of "High". This seems to populate for all records for that subject.

Also that's absolutely fine regarding the screenshot I will populate data using datelines for your reference in the future. I appreciate the feedback

[PaigeMiller]

Just exclude records that don't have the flag = 1 in your next step.

Or modify the SQL code. As I said, we need data as SAS data step code and not as screen captures to write anything but the simplest code.

[FreelanceReinh]

---

@smackerz1988 wrote:

Thanks for this but this but I need it to only populate for all records of the parameter that contains one record of "High". This seems to populate for all records for that subject.

---

Add the variable specifying the parameter (be it PARCAT1N or PARAMCD) to the GROUP BY clause:

group by usubjid, paramcd

[mkeintz · Accepted solution]

Assuming the data set file is already grouped by USUBJID/PARAMCD, then:

data want;
  do until (last.paramcd);
    set have;
    by usubjid paramcd notsorted;
    if a1ind='High' or a2ind='Change high' then ANL03FL='Y';
  end;
  do until (last.paramcd);
    set have;
    by usubjid paramcd notsorted;
    output;
  end;
run;