wwiecek / covid-vaccine-timeline

Modelling different COVID vaccine timelines

Vaccines at Velocity

This is a project repository for the following working paper:

Więcek, Witold, David Johnston, Tomas Dulka, Danny Toomey, and Enlli Lewis. ‘Vaccines at Velocity: Evaluating Potential Lives Saved by Earlier Vaccination in the COVID-19 Pandemic’. medRxiv, 20 June 2023. https://doi.org/10.1101/2023.06.16.23291442.

The last version of the paper is available as [Vaccines at Velocity.pdf].

How to re-generate the results from the paper from scratch

This is an orderly project. The directories are:

archive: Includes dependencies necessary to generate the model.

data: Contains the following data:

• excess_mortality: Fitted nimue models and pre-generated simulations, along with vaccine allocation information, for the fits to excess mortality.
• reported_deaths: Fitted nimue models and pre-generated simulations, along with vaccine allocation information, for the fits to reported COVID deaths.
• raw: Raw data used in modelling:
  • owid.rds: Our World In Data dataset used for vaccine allocation, downloaded 13-02-2022
  • excess_deaths.rds: Excess death estimates from the Economist, downloaded 13-02-2022
  • combined_data.Rds: Reported COVID deaths dataset, downloaded 13-02-2022
  • vaccine_agreements.rds, vaccine_doses_by_manufacturer.rds, who_vacc.rds, who_vacc_meta.rds: Other vaccination datasets, downloaded 13-02-2022
  • worldsf.Rds: World map sf used in the plotting, downloaded 20-04-2022 from https://datahub.io/core/geo-countries/r/countries.geojson

docs: Includes figures, tables, and legends used in the preprint.

figure: Includes the figures used in the preprint and report.Rmd, which can be used to generate the figures.

src: Contains structured orderly directories which generate the models used in the preprint. A seperate README and vignette are included in the src directory with directions to add counterfactual scenarios.

The purpose of this repository is to allow others to replicate the analysis used in our preprint and replicate the generation of the preprint in its entirety.

Installation

    git clone https://github.com/wwiecek/covid-vaccine-timeline.git
    cd covid-vaccine-timeline
    open covid-vaccine-impact-orderly.Rproj

Usage

The models used in the preprint can be generated from scratch by running run-fits.R. This operation is very resource intensive and may take several hours to complete.

Counterfactual scenarios can be generated by running run-counterfactuals.R. run-fits.R must be ran before running run-counterfactuals.R to generate the necessary dependencies. This operation is also resource intensive and takes approximately 1 hour to complete.

The preprint can be generated locally by running orderly::orderly_run("preprint"). This will create a PDF of the preprint at draft/preprint/[orderly id]/_book/_main.pdf. The draft directory is not included in this repository and will be generated automatically by orderly.

Common errors

Error while running run-counterfactuals.R:

Error in `map()`:
ℹ In index: 2.
Caused by error:
! vector memory exhausted (limit reached?)

Solution: Increase the amount of memory that can be allocated to the program. The amount of memory you want to increase the allocation to is up to you, though the authors recommend setting a large limit such as 50gb to avoid this error. This will not impact how R operates outside of the current environment. The following commands are given with the assumption the user wants to increase the memory to 50gb and can be adjusted to a lower size if desired.

• On Windows, run memory.limit(size=50000)
• On Mac, the solution is a bit more involved. Enter the following commands to create an R environment file (.Renviron) and add a maximum memory allocation argument (R_MAX_VSIZE):
  1. Navigate to the Terminal tab in RStudio (next to the Console in the top left).
  2. Enter touch .Renviron to create the environment file.
  3. Enter open .Renviron to open the environment file in a terminal window.
  4. Type R_MAX_VSIZE= 50Gb to set the memory allocation to 50gb and save the file.
  5. You can verify that the file was correctly saved by re-entering open .Renviron and checking for your changes.