Date:         Sat, 16 Aug 1997 18:09:45 +-300
Reply-To:     001 <jucha@NETVISION.NET.IL>
Sender:       "SAS(r) Discussion" <SAS-L@UGA.CC.UGA.EDU>
From:         001 <jucha@NETVISION.NET.IL>
Subject:      Re: Crack this
Content-Type: text/plain; charset="us-ascii"

Henrik wrote:
 
>>> I have a number of patients with a number of event dates. I would like
>>> to create datasets with patients who have more than one event at the
>>> same date, more than one event within 1, 2, 3, 4, 5, 6, 7 days.
 
>>> My dataset looks something like this:
 
>>> PATID   EVENTDAT
>>> 1       970303
>>> 1       970303
>>> 1       970405
>>> 2       970205
>>> 3       960725
>>> 3       960830
>>> 3       960902
 
>>> Maybe I can use the lag function but I am not sure how to code for this
>>> problem.
>>> Any suggestions?
 
Dear Henrik,
  I suggest you to try the following:
 
data a;
input EVENTDAT yymmdd6. PATID;
cards;
970303       1
970303       1
970405       1
970205       2
960725       3
960830       3
960902       3
;
run;
proc sort data=a;by patid eventdat;run;
data a0(keep=patid) a3(keep=patid) out(keep=patid);
  set a;
  by patid;
    retain mindif;
    ddat=dif(eventdat);
    if first.patid then do;
       ddat=.;
       mindif=.;
    end;
    mindif=min(ddat,mindif);
    if last.patid;
    if mindif=0 then output a0;
    if mindif=3 then output a3;
    if mindif=. or 7<mindif then output out;
run;
 
Elik Jucha
Statistics