retrorisk: a tool for estimating association between epidemics and diseases of uncertain aetiology

[adamkucharski]

retrorisk: a tool for estimating association between epidemics and diseases of uncertain aetiology

Description

The package would provide tools for retrospective analysis of the association between risk of infection during a recent epidemic and recent patterns of disease, the cause of which is currently have uncertain. For example, following Zika epidemics, cases of Guillain-Barre Syndrome (GBS) peaked shortly afterwards, whereas cases of microcephaly peaked much later, reflecting risk of Zika exposure in early pregnancy.

Estimation of risk can therefore be done by combining a statistical model of risk with four pieces of information: timeseries of epidemic cases, final size of epidemic, timeseries of the disease of uncertain aetiology; and likely relationship between infection and outcome (e.g. typical time from infection to GBS for other viruses).

Target audience

• typical end-users: organisations and agencies investigating diseases of uncertain aetiology

• potential contributors: field epidemiologists and researcher groups who have previously focused on this question

• key collaborators: Similar to potential contributors.

Interoperability

• inputs: data.frame (or tibble) with timeseries of epidemic cases and timeseries of the disease of uncertain aetiology

• outputs: data.frame with estimation of risk and comparison of statistical evidence for association

• imports:

  • epiparameter
  • finalsize

• used by:

  • scenario

• related projects

  • <*a package doing related things *>

Usage

The code below illustrates a typical use of the package, using fictitious code and outputs if needed:

library(retrorisk)

# Use case 1: 
epidemic_size <- final_size(r0=2)

combined_data |>
  estimate_risk(., epidemic_size, infection_to_outcome_process_model) |>
  summarise_results()