I’m trying to create an analytic plan to populate the table below:
| Number of clients | Baseline (Entry into the Program) | Follow-up Average |
| | No hospitalizations | 1+ hospitalizations | No hospitalizations | 1+ hospitalizations |
| | N | % | N | % | N | % | N | % |
Year of Follow-up | | | | | | | | | |
2008 | | | | | | | | | |
2009 | | | | | | | | | |
2010 | | | | | | | | | |
2011 | | | | | | | | | |
2012 | | | | | | | | | |
2013 | | | | | | | | | |
The variables that I have are:
ID – an Identifier that separates each case
EventCode – The type of Event
EventCount – A Count of all events
EventDate – The date when the event occurred
EnrollmentDate – The date when the client was enrolled
Here is my plan so far:
Step 1 – Subset only “Hospitalizations” using the EventCode Variable.
Step 2 – Create a BaselineEventDate variable by using:
proc sort data= Events; by ID EventDate;
data Events2; set Events;
by childid AsthmaEventDate;
if first.childid then BaselineEventDate = AsthmaEventDate;
retain
BaselineEventDate
Run;
The issue I see with this approach is what happens if this person did not have a Hospitalizations at baseline but did have one at follow up? How should we approach this?