Extract observations prior to one value of the indicator (dementia)

[BochengJing]

Hi SAS users, 

This may seem to be an easy and obvious programming but I just couldn't figure it out. I need your help. 

In my dementia data, I have patients with repeated records throughout the years. In some years, patients were diagnosed with dementia while in other years, patients did not. My data looks like this:  

Patient	Year	Dementia
A	2010	0
A	2012	0
A	2014	0
A	2016	1
A	2018	1
B	2010	0
B	2012	1
B	2014	0
B	2016	1
B	2018	0
C	2010	1
C	2012	0
C	2014	0
C	2016	1
C	2018	1

What I want is for each patient, all their the dementia=0 observations prior to the dementia=1 diagnosis. In another words, if a patient has been diagnosed with dementia (dementia=1), I don't care what happened after (once you are diagnosed, you are identified as dementia forever) -- I want to extract all the dementia=0 cases prior to their dementia=1 diagnosis. So the table I want looks like this: 

Patient	Year	Dementia
A	2010	0
A	2012	0
A	2014	0
B	2010	0

Can anyone help me with this coding? Thank you.

[PeterClemmensen]

data have;
input Patient $ Year Dementia;
datalines;
A 2010 0
A 2012 0
A 2014 0
A 2016 1
A 2018 1
B 2010 0
B 2012 1
B 2014 0
B 2016 1
B 2018 0
C 2010 1
C 2012 0
C 2014 0
C 2016 1
C 2018 1
;

data want;
   set have;
   by patient;
   if first.patient then _iorc_ = 0;
   if Dementia then _iorc_ = 1;
   if _iorc_ = 0;
run;

Result:

Patient Year  Dementia
A       2010  0
A       2012  0
A       2014  0
B       2010  0