Estimate of the hazard of death associated with infection by SARS-CoV-2 VOC 202012/01 relative to preexisting variants.
This repo includes an anonymised data set (in ./dataset/) which allows partial replication of the analyses in our preprint. This
is a qs file, which can be read using the qs R package.
Note that the anonymised data set randomises ages within 5-year age bands, and caps the maximum age at 104 for purposes of not publishing personally identifiable information. For this reason, results obtained using this data set will not be exactly equal to those found in the paper, which analyses the non-public data. (This just means we have the exact age in years for each person tested—we have never had access to things like names and addresses!)
The FINALID column in the anonymised data set is also randomised, so please be aware that this will not link correctly to other datasets held by Public Health England. This is again in order to minimise any risk of disclosure of personal information.
The code is designed to operate either on non-public data supplied by PHE, or on the anonymised data set packaged with the repo.
Anonymised data sets live in the ./dataset/ folder.
survival.R is the main analysis file for the survival analyses. Calls to the R function complete_data can be replaced with
calls to the R function reduced_data to load an anonymised data set. All output goes to the ./output/ folder which you will
need to create in the project directory.
If you either can't read qs files, or cannot download the anonymised data from this repository (possibly because I have run out of LFS storage space on Github), or just prefer to work with CSV format data, the anonymised data are also available to download in CSV format from https://zenodo.org/record/5105921.