02-27-2017 04:29 PM - edited 02-27-2017 04:29 PM
I'm looking to model the random effects of a nested three level hospital and hospital collaboratives on patient mortality.
The structure is: Patients nested within hospitals and Hospitals are nested within collaboratives . My concern is that a hospital can be in multiple collaboratives because collaboratives classification is not mutually exclusive. For example there are six collaboratives groups: A, ,B, C, D, E, F and a hospital can be in a collaborative A and also collaborative B.
Should I code collaborative groups as dummy variables A through F (0,1) and use that as nesting variables for hospitals?
I couldn't find any clear literature that addresses my issue. Any help or guidance to relevant literature would be appreciated.
I will be using SAS 9.4.
Thanks in advance.