topic Re: Need Help in SAS Programming

Need Help

Guptashwe — Tue, 15 Jan 2019 13:31:55 GMT

Cough is identified using ICD-9/10 codes from the claims databases. Whereas, there is no ICD-9/10
code available to define the chronic stage of cough. The objective of the problem is to identify
patients who have been diagnosed with chronic cough between 2007 and 2009. These patients will
be then assessed for their demographic characteristics. The following algorithm is deigned to
identify chronic cough patients:
A patient having at least three inpatient/outpatient claims for cough within the time window of 120
days anytime, between 2007 and 2009.
The following dataset has patients with service dates for cough from 2007 and 2009. Identify
patients diagnosed with chronic cough between 2007 and 2009.

Re: Need Help

Kurt_Bremser — Tue, 15 Jan 2019 13:47:27 GMT

Please post your data in usable form. Pictures can NOT be used in SAS. See my footnotes.

Re: Need Help

Guptashwe — Tue, 15 Jan 2019 13:53:36 GMT

Re: Need Help

Guptashwe — Sun, 06 Jun 2021 09:04:31 GMT

i have attached the bat file.

Re: Need Help

Astounding — Tue, 15 Jan 2019 13:58:12 GMT

Assuming that your data already embodies much of the work that needs to be done:

The observations include only cough-related observations
The observations are already subset for the years 2007 through 2009
The observations are in order by ENROLID SVCDATE

All that remains is to identify patients that have 3 observations within a 120-day window. Here is an approach to do that:

data want;

id_counter=0;

chronic='N';

do until (last.enrolid);

set have;

by enrolid svcdate;

id_counter + 1;

if dif2(svcdate) <= 120 and id_counter >= 3 then chronic='Y';

end;

do until (last.enrolid);

set have;

by enrolid;

output;

end;

drop id_counter;

run;

The top loop processes observations for one ENROLID, setting CHRONIC to Y or N. The bottom loop processes the same observations, outputting them with the assigned value for CHRONIC.