bug: hospitalization underprediction with `hosp-treat`

[papsti]

the current hosp-treat implementation simply adds an accumulator compartment, $T$, for acute care admission to count the number of individuals that are treated in hospital. there is no other difference to the hosp model, which means $T$ over time should match cumulative acute care admissions exactly. this is not the case currently (can be seen in antivirals pipeline, for instance), so there is something wrong with the implementation.

• write unit tests comparing results from hosp and hosp-treat
• hunt the bug down in the hosp-treat implementation and fix

[papsti]

i suspect the problem is that the flow into $T$ was added also as a flow out of each origin compartment, which is not what we want for an accumulator compartment

[KevinHZhao]

EPACmodel/inst/models/hosp-treat/flows.csv

Lines 12 to 15 in d7500c8

	I_A,T,admission_to_A_R,per_capita,Epi,Epi,Epi,Age,Age,Null
	I_A,T,admission_to_A_CR,per_capita,Epi,Epi,Epi,Age,Age,Null
	I_A,T,admission_to_A_CD,per_capita,Epi,Epi,Epi,Age,Age,Null
	I_A,T,admission_to_A_D,per_capita,Epi,Epi,Epi,Age,Age,Null

In these rows of flows.csv, I think the type column should be "per_capita_inflow" instead of "per_capita". From my old macpan2 code, that should add the per capita flow to T without changing the value of I_A.

[papsti]

thanks, i think you're right! 🥳