A Package for Respiratory Disease Open-source Forecasting (PROF). PROF is an R package for fitting and forecasting infectious disease incidence. The package ingests publicly-available confirmed hospital admission data fits mechanistic models to the data, and provides short-term probabilistic forecasts. Currently, the package supports fitting and forecasting the individual and combined burden of influenza and COVID-19 at the state level. An S[I]2HR and SE[I]2HR models are used to fit the two pathogens and both use a flexible time-dependent transmission term. Once the User selects a state, and either one or both viruses, the PROF sequential fitting procedure determines the joint posterior distribution for the model parameters. The joint posterior distribution is then used with the model to generate location-specific probabilistic forecasts of the near-term number of hospital admissions. If both viruses are chosen, this procedure is done twice and the total hospital burden forecast is estimated by combining the trajectory profiles of each disease in multiple ways: random, ordered, and in-between.
This R package requires an R or RStudio installation and Fortran/GCC compilers.
> library(devtools)
> install_git(url="https://github.com/predsci/PROF")
To clone the User-friendly Graphical User Interface for PROF:
> git clone https://github.com/predsci/PROF-shiny-app.git
The PROF/prof_dev directory provides an example script that is a good starting point for using the package. This script relies on two other scripts in the same directory and hence we suggest to start as follows.
Create a 'test' directory
> mkdir test
navigate to the 'test' directory
> cd test
Copy all three scripts from PROF/prof_dev to the 'test' directory
> cp /path/to/PROF/prof_dev/*R .
Open R or R studio and the 'example.R' script
Read the script and execute it, preferably first line by line. Please note that you will likely need to edit directory paths in the script as appropriate.
For more documentation and examples please see: https://predsci.github.io/PROF/
The development of PROF is supported by the Council of State and Territorial Epidemiologists (CSTE) and the Centers for Disease Control and Prevention (CDC) through cooperative agreement number NU38OT000297, as part of the "Development of forecast, analytic, and visualization tools to improve outbreak response and support public health decision-making" project.
Michal Ben-Nun ([email protected])
James Turtle ([email protected])
Pete Riley ([email protected])