Check whether SDEC early discharge needs an 'eligibility' factor.

Similar to thrombolysis, where we have added in a randomized check to deal with the fact that people may be eligible in a simple sense (meet onset + scan criteria) but real-world complexity results in a significant percentage not being eligible. 

Could this apply to SDEC admission avoidance too? Significant enough to model? What about with therapy? 