topic Re: Diagnosis codes in SAS Programming

Diagnosis codes

RebeccaB_ — Tue, 08 Dec 2020 21:16:16 GMT

Hi,

I'm trying to flag all individuals with diagnosis codes E10, E11, and E13. The problem is, I would like this to include all values for that overarching diagnosis code. So for example, if there are diagnosis codes E11.17, E11.21, E11.34, E10.22, E13.49, I would like to flag any with E10, E11, or E13 to be flagged. My code is below and thanks in advance for any help!

data work.diabetes_control_num;
set work.diabetes_control;
array diag{3} $ diagnosis_1 diagnosis_2 diangosis_3;
array a1c_last{1} a1c2;
do i= 1 to 1;
if a1c_last{i} >= 9 then dm_control=0; else if a1c_last{i}=. then dm_control=.; else dm_control=1;
end;
do i=1 to 3;
if diag{i} = in ("E10", "E11", "E13") then diag_control=1; else diag_control=0;
end;
run;

Re: Diagnosis codes

ballardw — Tue, 08 Dec 2020 21:35:21 GMT

@RebeccaB_ wrote:

Hi,

I'm trying to flag all individuals with diagnosis codes E10, E11, and E13. The problem is, I would like this to include all values for that overarching diagnosis code. So for example, if there are diagnosis codes E11.17, E11.21, E11.34, E10.22, E13.49, I would like to flag any with E10, E11, or E13 to be flagged. My code is below and thanks in advance for any help!

data work.diabetes_control_num;
set work.diabetes_control;
array diag{3} $ diagnosis_1 diagnosis_2 diangosis_3;
array a1c_last{1} a1c2;
do i= 1 to 1;
if a1c_last{i} >= 9 then dm_control=0; else if a1c_last{i}=. then dm_control=.; else dm_control=1;
end;
do i=1 to 3;
if diag{i} = in ("E10", "E11", "E13") then diag_control=1; else diag_control=0;
end;
run;

Actually almost there. If the values are as you say you could use SCAN or SUBSTR functions to get the first letters from your diag codevalues.

if Scan( diag{i},1,'.') in ("E10", "E11", "E13") then diag_control=1; else diag_control=0;

if substr(diag{i},1,3) in ("E10", "E11","E13") then ...

Alternate

Diag_control = ( Scan( diag{i},1,'.') in ("E10", "E11", "E13") );

(Exercise for the interested reader for the similar SUBSTR version).

This assignment works because will assign a 1 for true and 0 for false for any logic comparison.

You don't want to use "= in (<value list>)" just the "in (<value list>)"

Re: Diagnosis codes

Reeza — Tue, 08 Dec 2020 22:25:25 GMT

Use the colon operator.

if diag{i} in: ("E10", "E11", "E13") then diag_control=1;
else diag_control=0;

@RebeccaB_ wrote:

Hi,

I'm trying to flag all individuals with diagnosis codes E10, E11, and E13. The problem is, I would like this to include all values for that overarching diagnosis code. So for example, if there are diagnosis codes E11.17, E11.21, E11.34, E10.22, E13.49, I would like to flag any with E10, E11, or E13 to be flagged. My code is below and thanks in advance for any help!

data work.diabetes_control_num;
set work.diabetes_control;
array diag{3} $ diagnosis_1 diagnosis_2 diangosis_3;
array a1c_last{1} a1c2;
do i= 1 to 1;
if a1c_last{i} >= 9 then dm_control=0; else if a1c_last{i}=. then dm_control=.; else dm_control=1;
end;
do i=1 to 3;
if diag{i} = in ("E10", "E11", "E13") then diag_control=1; else diag_control=0;
end;
run;

Re: Diagnosis codes

ballardw — Tue, 08 Dec 2020 22:38:10 GMT

@Reeza wrote:

Use the colon operator.
if diag{i} in: ("E10", "E11", "E13") then diag_control=1;
else diag_control=0;
@RebeccaB_ wrote:

Hi,

I'm trying to flag all individuals with diagnosis codes E10, E11, and E13. The problem is, I would like this to include all values for that overarching diagnosis code. So for example, if there are diagnosis codes E11.17, E11.21, E11.34, E10.22, E13.49, I would like to flag any with E10, E11, or E13 to be flagged. My code is below and thanks in advance for any help!

data work.diabetes_control_num;
set work.diabetes_control;
array diag{3} $ diagnosis_1 diagnosis_2 diangosis_3;
array a1c_last{1} a1c2;
do i= 1 to 1;
if a1c_last{i} >= 9 then dm_control=0; else if a1c_last{i}=. then dm_control=.; else dm_control=1;
end;
do i=1 to 3;
if diag{i} = in ("E10", "E11", "E13") then diag_control=1; else diag_control=0;
end;
run;

Might be an issue if codes of E100, E110 or E130 exist. May not be an issue but there are just so blinking many ICD codes out there ...

Re: Diagnosis codes

Reeza — Tue, 08 Dec 2020 22:53:46 GMT

True, but you could also just add the periods to the end in that case but it really depends on the coding. In a perfect world you have a lookup table that has the code in a hierarchical fashion you can just filter out.

Re: Diagnosis codes

RebeccaB_ — Wed, 09 Dec 2020 15:45:45 GMT

Thank you for the help! I've used both scan and substr but all observations of diag_control are still being flagged as "0", but I can visually see an observation that, for example, has diagnosis_1 = E11.21. Any suggestions in changing my code so that I can capture this?

CODE:

data work.diabetes_control_num;
set work.diabetes_control;
array diag{3} $ diagnosis_1 diagnosis_2 diangosis_3;
array a1c_last{1} a1c2;
do i= 1 to 1;
if a1c_last{i} >= 9 then dm_control=0; else if a1c_last{i}=. then dm_control=.; else dm_control=1;
end;
do i=1 to 3;
if Scan( diag{i},1,'.') in ("E10", "E11", "E13") then diag_control=1; else diag_control=0;
end;
run;

Re: Diagnosis codes

Reeza — Wed, 09 Dec 2020 16:41:10 GMT

Drop the else, otherwise it will always get reset to 0.