Projecting the transmission dynamics of SARS-CoV-2 through the post-pandemic period

This R code is the basis of the regression analysis presented in this manuscript, now available on medRxiv here.

The following datasets are included in this repo:

• Corona4PP_Nat.csv: Weekly percent testing positive for each coronavirus strain based on reports to the National Respiratory and Enteric Virus Surveillance System (NREVSS). This data (from Mar 2018 through Feb 2020) is publicly available on the CDC website here. Full data used in paper is available through a data use agreement with the CDC.
• ILINet.csv: Weekly data from the U.S. Outpatient Influenza-like Illness Surveillance Network (ILINet), including the weekly number and percent of patients presenting with ILI. This data is available through the FluView Interactive dashboard here.

The code is organized in the following chunks:

1. Prep: Download required packages and load in data.
2. Cleaning: Clean datasets and calculate weekly incidence proxy over the time period (percent of clinic visits for ILI multiplied by percent positive for each CoV strain).
3. Calculate R: Use the Wallinga-Teunis method to estimate effective reproduction numbers (3-week moving geometric mean).
4. Regression: Calculate depletion of susceptibles for each strain and perform regression.

Acknowledgements: Many thanks to te Beest, et al. The code to calculate the effective reproduction numbers was modified from the code for their paper "Driving factors of influenza transmission in the Netherlands" (AJE 2013).