Existing packages for modelling H5N1 human-to-human transmission

[jamesmbaazam]

Following up on #284, what tools are available for modelling H5N1 human-to-human transmission scenarios and how differently would we model/analyse H5N1 compared to mpox given that human-to-human transmission driven outbreaks are still limited? What are some potential questions or common tasks to expect in analysing H5N1?

[adamkucharski]

Thanks for starting this discussion - could be a nice follow up blog post with more specific examples? One of the major differences in transmission dynamics is the likely importance network structure for mpox vs age-stratified SIR-like dynamics for respiratory infection.

A few areas of ongoing work with our and other packages:

• Using {epichains} and H5 cluster sizes in US to estimate R, including a prior based on H5N1 in Indonesia (with prior extracted and defined using {epiparameter}).
• Using {epichains} to explore potential outbreak dynamics following importation (which you're obviously aware of!)
• Adapting {ringbp} to include wider pathogen scenarios. There is also a specific H5N1 branch with quick edits to customise, but for sustainability this needs to be incorporated into the main PR changes @joshwlambert has suggested. This would be good target for a dev day/sprint once people back from IDM I think.
• Longer-term potential pandemic scenarios (e.g. if becomes more transmissible). {epidemics} would be well placed to handle this, especially as we incorporate more capacity to handle age-specific hospitalisation/fatality risk etc. Now we're building in {odin} backend with @bahadzie , will be much easier to add model variants (e.g. with spillover introductions as well as SIR dynamics).
• With @LloydChapman, we've also been collating some library models for {odin.dust} and {mcstate}. I've got an initial script to reproduce estimation of epidemic size from early exported cases for COVID using these models - will push branch shortly.