Comparison of interventions against COVID-19 outbreaks in homeless shelters

This repository contains Approximate Bayesian Computation (ABC) code and simulation code for the analyses in 'Comparison of infection control strategies to reduce COVID-19 outbreaks in homeless shelters in the United States: a simulation study' [1]. The code implements a discrete-time stochastic SEIR simulation model of COVID-19 transmission in a closed environment (here a homeless shelter) with importation of infection from the local community. The model is fitted to data on numbers of PCR-positive and negative individuals from outbreaks in 5 homeless shelters in San Francisco, Boston and Seattle, and used to predict the impact of different intervention strategies on the probability of averting an outbreak over 30 days in a representative homeless shelter into which a single latently infected individual is introduced.

Prerequisites

• R version 4.0.0

• The following R packages are required to run the code:

  • mvnfast
  • actuar
  • ggplot2
  • reshape2
  • Hmisc
  • gsubfn
  • gridExtra
  • doParallel
  • abind

Data

All data required to run the code is available in the data subfolder.

Installing

Clone/download this project into a folder on your machine using the green button at the top right of this page.

Running the code

The required R packages can be installed by running the following line of code in R

> install.packages(c("mvnfast","actuar","ggplot2","reshape2","Hmisc","gsubfn","gridExtra","doParallel","abind"))

The model calibration can then be run in R by entering

> source("run_calibration.R")

at the command prompt, or by navigating to the downloaded code folder in a terminal window on Mac/Linux and entering

% Rscript run_calibration.R

or in Windows command line by entering

C:\>"C:\<path>\<to>\Rscript.exe" C:<path>\<to>\run_calibration.R

The intervention simulations and sensitivity analysis can be run in R similarly with

> source("run_interventions.R")
> source("run_sensitivity_analysis.R")

or via the command line (Mac/Linux) with

% Rscript run_interventions.R
% Rscript run_interventions.R

or from the Windows command line with

C:\>"C:\<path>\<to>\Rscript.exe" <path>\<to>\run_interventions.R
C:\>"C:\<path>\<to>\Rscript.exe" <path>\<to>\run_interventions.R

The SEIR transmission model is implemented in the COVID_homeless_functions.R script, and the fixed model parameters are set in set_nat_hist_pars.R. The ABC Sequential Monte Carlo (SMC) algorithm can be found in ABC_SMC.R.

Built With

• R version 4.0.0 (2020-04-24)

Author

• Lloyd Chapman: [email protected]

License

This project is licensed under the GNU General Public License v3.0 - see the LICENSE.txt file for details

Acknowledgments

The ABC SMC code adapts the code for case study 2 in [2] available here to enable fitting to both discrete and continuous parameters following the ABC model-selection algorithm described in [3,4].

References

1. Chapman LAC, Kushel M, Cox SN, Scarborough A, Cawley C, Nguyen T, Rodriguez-Barraquer I, Greenhouse B, Imbert E, Lo NC. Comparison of infection control strategies to reduce COVID-19 outbreaks in homeless shelters in the United States: a simulation study. medRxiv. 2020. doi:10.1101/2020.09.28.20203166v1

2. Minter A, Retkute R. Approximate Bayesian Computation for infectious disease modelling. Epidemics. 2019;29:100368. doi:10.1016/j.epidem.2019.100368

3. Toni T, Welch D, Strelkowa N, Ipsen A, Stumpf MPH. Approximate Bayesian computation scheme for parameter inference and model selection in dynamical systems. J R Soc Interface. 2009;6(31):187–202. doi:10.1098/rsif.2008.0172

4. Toni T, Stumpf MPH. Simulation-based model selection for dynamical systems in systems and population biology. Bioinformatics. 2009;26(1):104–10. doi:10.1093/bioinformatics/btp619