Why do we require a cause of death concept to mark patient as dead?

Continuing the discussion from Mark Patient as Dead:

person.cause_of_death is part of the data model. Why do we then require a “Cause of death” concept to be defined before a person can be marked as dead? If we create an observation for cause of death, could it be optional (i.e., if a concept hasn’t been defined, then don’t create an observation)?

Tickets like TRUNK-3269 suggest that we’re using observations for recording cause of death. Why is that?

Also, searching JIRA, it doesn’t look like we have a ticket for adding a person.cause_of_death_noncoded. I thought we had agreed to create this. Or did I dream it?

We have CIEL concept 162580. The object can use that by default (Cause of death, non-coded)

@slubwama just created such a ticket in the last few hours here: https://issues.openmrs.org/browse/TRUNK-5405

I thought that concept was only used by the UI to populate the choice list for causes of death.

If we’re also saving an obs when a patient is marked as dead, that seems misguided.