Identifying treatment regimens from clinical data

Asquared — Wed, 01 Aug 2018 15:51:50 GMT

I am trying to obtain different treatment regimens from my data for firstline treatment (i.e treatment received only during the first visit) and, also, the frequency(count) by which these different treatment regimens occur grouped by diagnosis in order to finalize one treatment regimen for each diagnosis. (Note- firstline treatment regimen can be a single drug/combination of drugs depending upon the patient).
Also, what would be an efficient way to graphically represent the duration of the overall treatment regimens given that each date contains either one or more than one observations per patient_ID.The dataset looks like this:

data have;
input id diagnosis $ trt_date mmddyy10. drug_code $;
format trt_date mmddyy10.;
datalines;
50 CAD 1/24/2010 A
50 CAD 1/24/2010 B
50 CAD 1/27/2010 A
50 CAD 1/27/2010 B
50 CAD 1/30/2010 A
50 CAD 1/30/2010 B
65 CAD 1/23/2010 B
65 CAD 1/24/2010 A
65 CAD 1/26/2010 A
65 CAD 1/28/2010 A
78 CHD 2/21/2010 A
78 CHD 2/21/2010 A
78 CHD 2/25/2010 A
78 CHD 2/25/2010 B
78 CHD 3/1/2010 A
78 CHD 3/1/2010 B
80 CAD 1/21/2010 B
80 CAD 1/31/2010 B
80 CAD 2/10/2010 B
85 CHD 6/19/2010 A
85 CHD 6/19/2010 B
85 CHD 6/25/2010 A
85 CHD 6/25/2010 B
85 CHD 7/1/2010 A
85 CHD 7/1/2010 B
89 CHD 12/19/2010 A
89 CHD 12/19/2010 B
89 CHD 12/22/2010 A
89 CHD 12/22/2010 B
89 CHD 12/25/2010 A
89 CHD 12/25/2010 B
90 CAD 1/11/2011 C
90 CAD 1/11/2011 D
90 CAD 1/14/2011 C
90 CAD 1/14/2011 D
90 CAD 1/17/2011 C
90 CAD 1/17/2011 D

;
run;

The resulting output would look something like this:

Diagnosis	FirstLine_Regimen	Frequency (Count)
CAD	AB	1
	B	2
	CD	1
CHD	AA	1
	AB	2

I read other threads that nearly related to my question but couldn't quite derive a working code based off them.

Help would be appreciated! Thanks.

Re: Identifying treatment regimens from clinical data

PeterClemmensen — Wed, 01 Aug 2018 05:15:12 GMT

Always a good idea to specify the form of your desired result. This code gives you all your firstline treatments

proc sql;
   create table firstline_trt as
   select * from have
   group by id, diagnosis 
   having(trt_date)=min(trt_date);
quit;

Re: Identifying treatment regimens from clinical data

Asquared — Thu, 02 Aug 2018 04:56:21 GMT

Hi @PeterClemmensen

Just updated the question and added the desired output table. It would look something like this.

Diagnosis	FirstLine_Regimen	Frequency (Count)
CAD	AB	1
	B	2
	CD	1
CHD	AA	1
	AB	2

Re: Identifying treatment regimens from clinical data

rajdesai — Sun, 10 Nov 2019 00:19:10 GMT

How do you create the frequency table?

topic Identifying treatment regimens from clinical data in SAS Programming

Identifying treatment regimens from clinical data

Re: Identifying treatment regimens from clinical data

Re: Identifying treatment regimens from clinical data

Re: Identifying treatment regimens from clinical data