R disease modeling

[kamayi585]

I’m Eric Noël Kamayirese from Rwanda. If anyone here could kindly share notes on R disease modeling and guidance on how to become familiar with it in a short time, I would greatly appreciate it. Thank you.

[avallecam]

Hi @kamayi585, there are some materials we usually point to learners:

At https://data.org/initiatives/epiverse/training-africa/course/, you will find:

• In "Phase 1: Foundational modules", you will find a section called "Epidemic and Mathematical Theory".
  • It contains videos, reading materials, assessments, and coding tutorials using {deSolve}.
• Then, in "Phase 2: Epidemiological data science", you will find the section "Scenario Modelling for Outbreak Analytics with R".
  • It contains tutorials using {epidemics}, which provide access to reusable compartmental structures as functions.
  • It prioritises outbreak-prone diseases with person-to-person transmission.

Alternative resources with theory details and practice that also use {deSolve} are:

• ebook: https://amanda-minter.github.io/intro-to-modelling/
• course: https://thegraphcourses.org/courses/introduction-to-infectious-disease-modelling/

One feature of {epidemics} is its interoperability with {odin}. The last one has a book with further documentation on how to use it, including methods for model fitting and inference using {monty}:

• example: https://epiverse-trace.github.io/howto/analyses/simulate_transmission/epidemics_odin_interop.html
• ebook: https://mrc-ide.github.io/odin-monty/

[avallecam]

cc'ing @Degoot-AM for further input

[kamayi585]

Thank you for the prompt feedback. However, I would like further clarification on how local or sub-national data can be used for disease modeling and forecasting specific to my region. Is this approach feasible at district or regional level, or is it intended to be applied only at the national level?

[kamayi585]

Thank you for the prompt feedback. However, I would like further clarification on how local or sub-national data can be used for disease modeling and forecasting specific to my region. Is this approach feasible at district or regional level, or is it intended to be applied only at the national level?

[avallecam]

Thanks for clarifying. I can say that yes, it is possible to contextualise the modelling to the subnational level. You can detail the population-specific parameters, like the age-stratified population size, the initial conditions, and the social contact matrix. You can double-check the study design details of the social contact surveys to assess how appropriate they are to fit your assumptions.

You can read further about this in the section we shared above: "Phase 2: Epidemiological data science" > "Scenario Modelling for Outbreak Analytics with R".