case statement

[tmm]

The setting up of arrays or variables does not work because I need all the elapsed times. I need to calculate all 30 columns to get elapsed times from all of them and then when I have the elapsed times of all then I take any that are 6 days to 31 days as a count of 1. They either have a count of 1 in the end meaning Yes they were readmitted within that 6 to 31 day period or 0 for no. But that comes later. I think I am just going to have to do it the way I am where each is time1, time2, time3 etc and then the 2nd case getting the actual amounts of times that elapse and then that is readm1, readm2, etc all the way to 29 and then I will have to either concatenate it all but having issues figuring out how to do that on numbers in SAS or maybe change the numbers to a var and concatenate. Not sure.

[Reeza]

So you're looking at 30/60/90 day readmits?

I still think arrays would come in handy for that.

Another method is to transpose the data so that each admit/discharge has its own record and then join on the same id where the difference between them is less than 30/60/90.  Not ideal but it works. 

And then you have the problem of having multiple visits within 30 days  

Admission/discharge data always makes my head hurt.

[Ksharp]

Why not at end keep only these new created variables. then use proc transpose.