topic Re: Compute summary variables in SAS in SAS Programming

Compute summary variables in SAS

Kimberly09 — Thu, 28 May 2020 22:17:17 GMT

Hi Community,

I hoping you could help me determine how to create and compute 2 new summary variables that capture the total number of contacts based on provider type (MD or Social worker (SW)). The 2 new variables need to be continuous. As you can see in the table below, study_id 1001 received a total of 8 contacts during the follow-up period: 6 MD & 2 SW. I would to write a SAS code to create and compute the 2 variables (MD_tot and SW_tot). Please help or point me in the right direction.

Thanks so much!

Data set: 2250 observations

Total # variables in data set: 138

Variables related to this question: (in table below)

Variables related to this question are coded as: Con1-Con8 (1 =MD, 2 =SW; binary)

New variables I'd like to create: count of total Con1-Con8 by provider contact. These will be continuous variables

Sample subset of data set

study_id	con_tot	con1	con2	con3	con4	con5	con6	con7	con8	MD_tot	SW_tot
1001	8	1	1	1	1	1	2	2	1	?	?
1002	2	1	2							?	?

Re: Compute summary variables in SAS

biopharma — Thu, 28 May 2020 22:37:18 GMT

Does this work for you?

data have ;
   input study_id con_tot con1 con2 con3 con4 con5 con6 con7 con8 ;
   cards ;
1001 8 1 1 1 1 1 2 2 1		
1002 2 1 2 . . . . . .		
;
run ;

data want ;
   set have ;
   array _con (*) con1 - con8 ;
   do _n_ = 1 to dim(_con) ;
      md_tot = sum(md_tot,(_con(_n_)=1)) ;
      sw_tot = sum(sw_tot,(_con(_n_)=2)) ;
   end ;
run ;

Re: Compute summary variables in SAS

ChrisNZ — Thu, 28 May 2020 23:00:39 GMT

Or:

data WANT;
   set HAVE ;
   MD_TOT = lengthn(compress(catt(of CON1-CON8),' .2'));
   SW_TOT = lengthn(compress(catt(of CON1-CON8),' .1'));  
run;

Re: Compute summary variables in SAS

Kimberly09 — Thu, 28 May 2020 23:45:26 GMT

The code works like a charm! Thanks so much for your help.

Re: Compute summary variables in SAS

Kimberly09 — Thu, 28 May 2020 23:49:15 GMT

Thanks so much! This works perfectly!