Hi all;
Kinda stumped on this and I'll see if I can explain it well
What I have are patient encounters with multiple catheter insertions and/or multiple surgeries-
So a patient encounter may be
PT_ENC_NUMBER CATH_TIME1 CATH_TIME2 SURG_START1 SURG_START2 SURG_END1 SURG_END2
123 01JAN17:12:00:00 05JAN17:12:00:00 01JAN17:10:00:00 07JAN17:13:00:00 01JAN17:13:00:00 07JAN17:17:00
So here Cath 1 was inserted in surg 1 but not Surgery2
Cath2 was not inserted in surgery1 or surgery2. So I know the total number of surgical insertions was 1-
All in all, its dynamic but let's say a patient can have up to 30 surgeries and 8 catheter insertions
I think I need a multi-dimension array but having some challenges overall;
Any advice appreciated.
Lawrence
You don't really show any desired result or explained it very clearly so I'll be guessing here.
I suspect that you want 3 separate arrays for the comparisons, one that would be the catheter times and the second the surgery start and the last the surgery end_end times.
Depending upon the desired result something like this may get you started. This assumes none of the surgery start/end pairs overlap and that the insertion occurs within a single interval and that all of your datetime variables are SAS valued datetimes. If they are character this won't work.
data example; set have; array c cath_time: ; array ss surg_start: ; array se surg_end; array ws {8}; /* this willl hold which surgery had an insertion number MUST match number of cath_time variables*/ do 1 = 1 to dim(ss); do j= 1 to dim(c); if ss[i] le c[j] le se[i] then do; /* found an insertion in the ith surgery for the jth cath*/ ws [j] = i; end; end; End; /* if you want a total number of cath insertions*/ NumInsert = n(of ws(*)); run;
Hi Lawrence,
My answer assumes that you have the start and end time for every surgery the patient had (aka. matching surg_startx and surg_endx).
data temp;
input PT_ENC_NUMBER @@;
input (CATH_TIME1 CATH_TIME2 SURG_START1 SURG_START2 SURG_END1 SURG_END2) ( :datetime.);
format CATH_TIME1 CATH_TIME2 SURG_START1 SURG_START2 SURG_END1 SURG_END2 datetime.;
datalines;
123 01JAN17:12:00:00 05JAN17:12:00:00 01JAN17:10:00:00 07JAN17:13:00:00 01JAN17:13:00:00 07JAN17:17:00
124 10JAN17:12:00:00 15JAN17:12:00:00 01JAN17:10:00:00 07JAN17:13:00:00 01JAN17:13:00:00 07JAN17:17:00
124 01JAN17:12:00:00 07JAN17:15:00:00 01JAN17:10:00:00 07JAN17:13:00:00 01JAN17:13:00:00 07JAN17:17:00
;
run;
data solution;
set temp;
array caths [*] cath: ;
array surg_s [*] surg_start: ;
array surg_e [*] surg_end: ;
total=0;
do i = 1 to dim(caths);
do j = 1 to dim(surg_s);
if surg_s[j]<=caths[i]<=surg_e[j] then total+1;
end;
end;
run;
You don't really show any desired result or explained it very clearly so I'll be guessing here.
I suspect that you want 3 separate arrays for the comparisons, one that would be the catheter times and the second the surgery start and the last the surgery end_end times.
Depending upon the desired result something like this may get you started. This assumes none of the surgery start/end pairs overlap and that the insertion occurs within a single interval and that all of your datetime variables are SAS valued datetimes. If they are character this won't work.
data example; set have; array c cath_time: ; array ss surg_start: ; array se surg_end; array ws {8}; /* this willl hold which surgery had an insertion number MUST match number of cath_time variables*/ do 1 = 1 to dim(ss); do j= 1 to dim(c); if ss[i] le c[j] le se[i] then do; /* found an insertion in the ith surgery for the jth cath*/ ws [j] = i; end; end; End; /* if you want a total number of cath insertions*/ NumInsert = n(of ws(*)); run;
Ballard;
As I said I having some trouble articulating what the results I was trying to get but you nailed 99.9% of the way there. I had to make a minor change but saved me a lot of hours spinning my wheels!
Thanks again-I love this forum!
Lawrence
Ok Ballardw I am back with a follow up question-
So I have expanded the array a bit and everything works pretty well except for the components in red. I can tell by manually comparing the data.
The expanded version not just looks if the catheter was placed during a surgery but also there is a binary flag to whether that surgery was under 4 hours. SO now I need surgeries whether catheter was placed in OR and another array to tell me if catheter was placed in OR and surgery was <4 hours. The array that calculates whether it was done in surgery works great (Thank you again)
I have tried different combinations to no avail.
data hairyarray;
retain cath_day SURG SURG_INSERT POD surg_under4hr SURG_INSERT4HR ED_STAY INPT_INSERTION;
merge toarray final_cath_transpo final_surg_transpo final_cath_transpo_placement;
by pat_enc_csn_id;
array c START_DT: ; /*ARRAY FOR THE CATH START TIMES*/
array ce END_DT: ; /*ARRAY FOR THE CATH END TIMES*/
array ss LOG_PAT_IN_ROOM_TIME: ; /*ARRAY FOR THE SURG START TIMES*/
array se LOG_PAT_OUT_OF_ROOM_TIME: ; /*ARRAY FOR SURG END TIMES*/
array Surgfour surg_under4hr: ; /*ARRAY FOR THE SURG FOR UNDER 4 HOURS*/
array ws {&finalcath} ; /*ARRAY TO DETERMINE IF CATH PLACED IN OR*/
array podx {&finalcath} ; /*ARRAY TO DETERMINE IF CATH REMOVED POD1*/
array surgx{&finalsurgc} ;/*ARRAY TO COUNTS NUMBER OF SURGERIES*/
array SURGfouri{&finalcath} ;/*ARRAY TO DETERMINE IF CATH PLACED IN OR AND SURG UNDER 4 HOURS*/
do i = 1 to dim(ss);
do j= 1 to dim(c);
if ss[i] le c[j] le se[i] then do;
ws [j] = i;
if ws [j] = &finalsurgc then ws [j]=.; /*DELETES INCORRECT ENTRIES*/
if ws [j]=i and Surgfour[i]^=. then SURGfouri[j]=i;
if ws [j]^=. and intck('dtday', se[i],ce[j])<2 then podx[j]=i; /*THIS GETS WHETHER IT WAS POD1 OR NO*/
if podx[j]=&finalsurgc-1 then podx[j]=.; /*DELETES INCORRECT ENTRIES*/
end;
if ss[i] ^=. then surgx[i]=1; /*COUNTS THE SURGERIES*/
if SURGfouri[j]>n(of surgx(*)) then SURGfouri[j]=.;
end;
end;
; /*TOTALS NUMBEROF SURGICAL INSERTIONS*/
POD = n(of PODx(*)); /*TOTALS NUMBER OF POD1 REMOVALS*/
SURG=n(of surgx(*)); /*TOTALS NUMBER OF SURGERIES*/
SURG_INSERT = n(of ws(*)); /*TOTALS NUMBER OF SURGICAL INSERTIONS*/
SURG_INSERT4HR=n(of SURGfouri(*)); /*TOTALS NUMBER OF SURGICAL INSERTIONS <4 HRS*/
run;
Thanks for your time again.
Lawrence
(ps If I am not making sense I apologize it's late)
It's finally time to hack! Remember to visit the SAS Hacker's Hub regularly for news and updates.
Learn how use the CAT functions in SAS to join values from multiple variables into a single value.
Find more tutorials on the SAS Users YouTube channel.
Ready to level-up your skills? Choose your own adventure.