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).

The Cauchemez et al (2016) analysis of ZIka in French Polynesia initially used this approach to identify first trimester as the highest risk period from pregnant women post-Zika infection, based on available data.

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()