Open daniel-klein opened 11 months ago
Lacking cause of death at the People level, we are getting proliferation of States like ti_dead
within each Disease module.
We should definitely do this, but recognize that it's hard to implement consistently. You have TB and HIV and you die. Was your cause of death TB or HIV? Well, it was both. There are best practices for doing this (see e.g. the Global Burden of Disease approach), but they're complicated.
Can we develop a Comorbidity Matrix similar to a Covariance Matrix where we have random draws to indicate the Contribution to Cause of Death (Scenario A){HIV: 100%, TB: 0%} OR (Scenario B){HIV:70%: TB:30%}?
Considering we have demographics as well as multiple disease modules, it will be useful to record which module (or demographics) was responsible for the death of each agent during the course of a simulation.