10-06-2015 12:06 PM
I'm analyzing emergency room utilization rates at several sites and want to do a multiple comparison test. I have the raw data (ER visits and Enrollees) and was wondering if someone could suggest a way to analyze them. I'm thinking it would be a GenMod or Glimmix solution.
10-06-2015 04:04 PM
I adapted the code from here
data EDRate; input v e Group$; ln = log(e); datalines; 23 155 Group1 29 276 Group2 54 550 Group3 ; proc genmod data=EDRate; class Group; model v = Group / dist=poisson link=log offset=ln; lsmeans Group / ilink cl; run;
where v are the number of ED visits and e are the number of enrollees. Does this look like an appropriate way to analyze the data? Any helpful hints would be greatly appreciated.