calculate new mean from baseline and multiple visits using LAG

[jenim514]

I'm trying to calculate the change in mean from baseline to each visit (vist1 to vist2; vist1 to visti3; visit1 to visit4...)

I tried this using lag, but results don't seem correct.  It is just using the lag between observations rather than holding visit 1 as the LAG.  Help is appreciated

Starting:

visit	mean
1	2.41
2	2.47
3	2.53
4	2.54

want something like this:

visit	mean	change
1	2.41	.
2	2.47	2.44
3	2.53	2.47
4	2.54	2.48

I started with this code:

data want;

set have;

lag_mean=lag(mean);
change=mean-lag_mean;
run;

[Reeza]

Try an approach using RETAIN instead.

[Ksharp · Accepted solution]

data have;
input visit	mean;
cards;
1	2.41
2	2.47
3	2.53
4	2.54
;
run;

data want;
set have;
sum+mean;
change=sum/_n_;
if _n_=1 then change=.;
drop sum;
run;