topic SAS code for counting cases/control for unique id_idchem combinations in SAS Studio

SAS code for counting cases/control for unique id_idchem combinations

ak2011 — Wed, 01 Jan 2020 15:49:22 GMT

Hi,

Please, could someone help me with the SAS code to verify the count(frequency) of ca case ca cont pop cont for each unique/distinct id and idchem combination. I have 29 obs, 3 pollutants idchem 990005=cla_exp, 990021 = bio_exp and 210701 = amo_exp. I need to count the number of the lung variables associated with each unique id_idchem combination. For obs 2 (Osaa13-idchem 210701) has 1 lung ca case. Same with obs 4 so I expect SAS to count obs 2 & 4 as frequency of 1 since they have the same id_idchem combination.

Thus, I would like the overall freq of unique id_idchem by lung for 210701 to be 2/8 for ca case (for example).

My SAS output Tables 1 & 2 are attached. The SAS code and log are found below.

1 OPTIONS NONOTES NOSTIMER NOSOURCE NOSYNTAXCHECK;

73 data mg1b; set mgp(obs=29);

NOTE: There were 29 observations read from the data set WORK.MGP.

NOTE: The data set WORK.MG1B has 29 observations and 4 variables.

NOTE: DATA statement used (Total process time):

real time 0.01 seconds

cpu time 0.01 seconds

74 proc print data=mg1b;

75 Title " Table 1: Exposure to 3 pollutants";

NOTE: There were 29 observations read from the data set WORK.MG1B.

NOTE: PROCEDURE PRINT used (Total process time):

real time 0.27 seconds

cpu time 0.27 seconds

77 proc freq data=mg1b;

78 tables idchem*lung;

79 Title" Table 2: Frequency: id_idchem vs case/control for 3 pollutants "; run;

NOTE: There were 29 observations read from the data set WORK.MG1B.

NOTE: PROCEDURE FREQ used (Total process time):

real time 0.21 seconds

cpu time 0.20 seconds

81 OPTIONS NONOTES NOSTIMER NOSOURCE NOSYNTAXCHECK;

Thanks very much.

ak.

Re: SAS code for counting cases/control for unique id_idchem combinations

ed_sas_member — Wed, 01 Jan 2020 17:16:30 GMT

Hi @ak2011

Happy new year!

I suggest that you run a PROC SORT with the NODUPKEY option before running the PROC FREQ:

proc sort data = env out = env_distinct nodupkey;
	by lung id_idchem;
run;
proc freq data = env_distinct;
	table lung * idchem / nocol norow nocol nopercent;
run;

Does that give you the expected results?

Re: SAS code for counting cases/control for unique id_idchem combinations

mkeintz — Wed, 01 Jan 2020 18:20:27 GMT

proc sort with NODUPKEY is useful, but depending on the size of the data set, can be expensive. In this case, I'd suggest two proc freqs. The first produces a data set FREQS of frequencies of IDCHEM*LUNG (i.e. one obs for each combination with 2 additional variables COUNT and PERCENT). Then submit the new dataset to a proc freq with the same tables specification. This will yield counts of non-empty combinations:

proc freq data =mgp noprint;
  tables idchem * lung /out=freqs;
run;
proc freq data=freqs;
  tables idchem * lung / norow nocol;
run;

The advantage here is that proc freq doesn't need a sorted data set.

Re: SAS code for counting cases/control for unique id_idchem combinations

ak2011 — Thu, 02 Jan 2020 11:32:09 GMT

Thanks very much. It works!
ak.