topic macros in SAS Programming

macros

manya92 — Mon, 02 Jul 2018 15:26:06 GMT

%macro demo;
%let ccnames=Myocardial infarction*Congestive heart failure*Peripheral vascular disease*Cerebrovascular disease*Dementia*Chronic pulmonary disease*Rheumatic disease*Peptic ulcer disease*Mild liver disease*Diabetes without chronic complication*Diabetes with chronic complication*Hemiplegia or paraplegia*Renal disease*Any malignancy*Moderate or severe liver disease*Metastatic solid tumor*AIDS/HIV;
%let cc1 = %str('I21','I22','I252','410','412');
%let cc2 = %str('I099','I110','I130','I132','I255','I420','I425','I426','I427','I428','I429','I43','I50','P290','39891','40201','40211','40291','40401','40403','40411','40413','40491','40493','4254','4255','4256','4257','4258','4259','428');
%let cc3 = %str('I70','I71','I731','I738','I739','I771','I790','I792','K551','K558','K559','Z958','Z959','0930','4373','440','441','4431','4432','4433','4434','4435','4436','4437','4438','4439','4471','5571','5579','V434');
%do i=1 %to 17;
%let ccname=%scan(&ccnames., &i., *);
proc sql;
create table _15_cohort_cc&i. as
select distinct enrolid, "&ccname." AS cc&i.
from derived._15_diagtable a inner join derived._15_firstproc b
on a.enrolid=b.enrolid and a.code in: (&&cc&i.) and b.first-180<=a.svcdate<=b.first
order by enrolid;
quit;
%end;

what is this code doing ? i do not understand macros that well, any help is appreciated

Re: macros

RW9 — Mon, 02 Jul 2018 15:38:57 GMT

It is creating a set of macro lists E.g. (and note how I am using the code window - its the {i} above post):

%let cc1 = %str('I21','I22','I252','410','412');

And then a bit of code with those lists.

%do i=1 %to 17;
%let ccname=%scan(&ccnames., &i., *);
proc sql;
create table _15_cohort_cc&i. as
select distinct enrolid, "&ccname." AS cc&i.

This is basically looping over each of the ccnames. list and creating a dataset for each one.

My opinion is this is a very bad coding and data modelling setup, and I personally would delete it.

Put data in datasets then work with Base SAS - which is the programming language - to manipulate the data. This is how programming should work. Macros do have a use, but doing data processing and such like it is not it.

Re: macros

manya92 — Mon, 02 Jul 2018 15:43:31 GMT

If i have to look for 139 codes from 001-139 which can have sub categories , how can i achieve this by simple SAS coding ? or using %str. Also my code list includes ICD9 codes which are not in sequence , for Eg.

254.1, 255.8, 323.x etc.

and i want to look for both 001-139 and the second list of random not sequenced codes using simple SAS coding ?

Re: macros

Kurt_Bremser — Mon, 02 Jul 2018 16:16:54 GMT

What would you be doing with the split datasets? This will determine a probable by-group processing.

Re: macros

manya92 — Mon, 02 Jul 2018 16:25:08 GMT

This is the exclusion crtieria

Any patient who had a serious or opportunistic infection within 30 days prior to the index date.
1.1 Infections resulting in hospitalization: any disease/condition listed in the final outcome definition that was coded in any of the 15 diagnosis fields of an inpatient admission claim.

Or

Infections resulting in administration of intravenous antibiotics: any disease/condition listed in the final outcome definition that was coded in any of the 9 diagnosis fields of an outpatient services claim and which was treated with an intravenous antibiotic as defined by specific Healthcare Common Procedure Coding System code or a National Drug code. Intravenous antibiotic therapy could not be more than 7 days before or 7 days after the infection diagnosis.

001
002
003
004
005
006
007
008
009
010-018 Tuberculosis*
010
011
012
013
014
015
016
017
018
020-027 Zoonotic bacterial diseases
020
021
022
023
024
025
026
027
030-041 Other bacterial diseases
030
031
032
033
034
035
036
037
038
039
040
041
042 Human Immunodeficiency Virus
042
045-049 Poliomyelitis/other non-arthropod viral diseases of CNS*
045
046
047
048
049
050-057 Viral diseases accompanied by exanthem
050
051
052
053
054
055
056
057
060-066 Arthropod-borne viral diseases
060
061
062
063
064
065
066
070-079 Other diseases due to viruses & chlamydiae
070
071
072
073
074
075
076
077
078
079
080-088 Rickettsioses & other arthropod-borne diseases
080
081
082
083
084
085
086
087
088
090-099 Syphilis & other venereal diseases*
090
091
092
093
094
095
096
097
098
099
100-104 Other spirochetal diseases
100
101
102
103
104
110-118 Mycoses*
110
111
112
114
115
116
117
118
120-129 Helminthiases
120
121
122
123
124
125
126
127
128
129
130-136 Other infectious & parasitic diseases
130
131
132
133
134
135
136
137-139 Late effects of infectious & parasitic diseases
137
138
139

254.1
255.8
323.x
324.x
357.0
380.1
382.0
382.9
383.x
384.0
384.1
421.x
422.x
424.9
447.2
457.2
460
461.x
462
463
464.x
465.x
466.x
472
473.x
474.0
475
476.x
480
481
482
483
484
485
486
487.1
487.8
490
491.2
510.x
511.1
513.x
519.1
519.2
522.5
522.7
523.3
528.3
528.5
529.0
540-542
550.0
551.x
566
567.x
569.5
572
572.0
572.1
574.0
574.1
574.3
574.4
575.0
575.1
576.1
590.x
595.x
597.0
597.8
599.0
603.1
604.x
607.1
607.2
614.x
615.0
615.9
616.x
680.x
681.x
682.x
684
685.0
686.x
711.0
711.9
728.0
729.4
730.x
785.4
788.7
790.7
790.8
996.6
998.5
999.3

I need to look for all these codes as they are the for infection conditions requiring hospitalization or IV antibiotics.

how can i do this ?

I have my cohort with one row per patient called cohort1

Re: macros

RW9 — Mon, 02 Jul 2018 16:32:22 GMT

I have given you the code on your other post. First, you create a dataset of the data - this being the codes from the exclusion criteri - often people like to put these into a spreadsheet and read that in. But you can do it in a datastep:

data codes;
  input code $;
datalines;
001
002
003
004
005
006
007
008
;
run;

Next we uitilise that code to find out from a dataset demo, which of those pt's have this code:

proc sql;
  create table exclusion_pts as
  select pt
  from   demo 
  where  code in (select code from codes);
quit;

Very simple Base SAS, with data all in datasets with formats and such like and all the Base SAS functionality available to process it. You could even improve on this to expand the list of codes automatically, maybe:

data codes (drop=i);
  input code $;
  if index(code,"-") then do;
    do i=input(scan(code,1,"-"),best.) to input(scan(code,2,"-"),best.);
      code=put(i,z3.);
      output;
    end;
  end;
datalines;
001-007
;
run;

The above does the ones with a hyphen in for example.