Intervention: Vaccination

[ptheywood]

• New agent variable (part of the disease model), float/double, degree of immunity, [0,1]
• On vaccination, immunity increases
• Over time, immunity wanes, computed per individual?
• Secondary vaccinations will increase the level of immunity to a different value?
• Vaccination occurs after some triggers
  • Initially just time based, with a rate of vaccination?
  • Start with different age groups?
  • Maybe based on occupation network.

ODE struggle with multiple vaccination - one of the appeals of ABM.

[ptheywood]

Individuals are vaccinated during an intervention, which may be:

• Vaccination of a specific individual
• Vaccination of people per age group (as a fraction of the age group)

When an individual has been vaccinated, their status is updated and an event is added to the event list for them N days later for the outcome of vaccination to be decided.
On the day of the vaccine taking effect, the individuals status is updated, and subject to the vaccine efficacy (using a single RNG value for all 3 levels of efficacy).
Waning of the vaccine is then added to the individuals event list, for the expected time of waning (current time + vaccine effect duration - vaccine time to protection)

---

The use of vaccination only appear to be possible through the R and python interfaces to the reference model and not the C implementation's IO parsing / cli.

In the ``multi_strain_vaccinate` example, a simulation is constructed, a vaccine type defined, and a vaccine schedule created. The simulation is executed for 30 steps one at a time, prior to lockdown being enabled, and individuals vaccinated every day for 70 days, before simulation resuming without lockdown for another 40 days

I.e. there is no reference input file configuration option for this

---

As we would like to demonstrate the benefits of ABMs compared to compartemental models, I'll not reproduce this exactly in this implementations.

I'll also have to devise an input file-based approach for scheduling the vaccination, which will only support a limited vaccination schedule (which is fine as this implementation is currently a proof of concept/demonstrator of using FLAME GPU 2 for this type of model).

This will be a quite restrictive input format for now, and either need to be another cli parameter, a relative path within the existing parameters file, or very restrictive additional parameters in the additional parameters file.