/basic-reproduction-number-covid19

This repository contains a analysis of the basic reproduction number of COVID-19 in Spain

Primary LanguageR

Basic reproduction number (R0) of the COVID-19 outbreak in Spain

Code repository with the aim of having an estimate of the evolution of the COVID-19 outbreak in the Spanish regions.

The code is developed entirely in R. The drake package is used to manage the workflow. This analysis can be run by cloning the repository and running the make.R file.

Disclaimer (!)

This analysis is developed by a layman and has the sole objective of having a minimum idea of the potential evolution of the outbreak in Spain. Under no circumstances this is intended to be a reliable study with which to take any health decision. If you want to figure out how experts are modelling this, I strongly recommend taking a look at the excellent work that the London School of Hygiene and Tropical Medicine is doing.

Assumptions

The analysis focuses on the estimation of the basic reproduction number for the Spanish regions based on the confirmed cases by region. The EpiEstim package is used for this purporse. In particular, this package has a method (uncertain_si) in order to account for uncertainty on the serial interval distribution. Why was this method chosen?

  1. It’s a new virus so a lot of research will be done in the coming months. There’s no absolute evidence about the serial interval distribution.
  2. Although many epidemiological models for viruses with similar behaviour are modelled with a serial interval under a Gamma or Weibull distributions, this doesn’t seem to be entirely clear in this case.
  3. It’s true, as discussed in the article linked in point 2, that the virus can have a normally distributed serial interval and therefore not restricted to positive values. In this case, EpiEstim doesn’t consider distributions with real domain for the time being, so in this analysis the following normal distribution is considered:

  • equation truncated at equation

  • equation truncated at equation

Reproducible analysis with Docker

To run the analysis, a Docker image is provided that reproduces the results using the most updated information from the Instituto de Salud Carlos III.

Note: Depending on your configuration you might need to run the following commands with root privileges (with sudo).

First install docker and clone this repository:

git clone https://github.com/jlopezper/basic-reproduction-number-covid19.git

Navigate to the cloned repository on your own machine, and launch the container:

cd /path/to/repo
docker-compose up -d

Enter the container and run the make.R file:

docker exec -it basic-reproduction-number-covid19_analysis_1 bash
Rscript make.R

When the process ends, exit the container:

exit
docker-compose down

Finally, you will find the results within the plots directory!

Image

Useful resources