O2COV2 analysis repository

This repository contains the R code used for the O2COV2 study analyses. The analysis is organised as a reproducible pipeline using the targets R package.

Due to data protection constraints, this repository is not fully runnable as-is.

The original pipeline starts from raw REDCap data, which cannot be shared publicly because they contain sensitive information. Instead, we provide preprocessed, analysis-ready datasets limited to the variables used in the study.

These datasets correspond to intermediate objects of the pipeline, allowing the main analyses to be reproduced without access to the raw data.

Clean datasets are publicly available on Zenodo:

https://doi.org/10.5281/zenodo.10480917

How to reproduce the analyses

The full pipeline defined in _targets.R documents all data processing and analysis steps.

To reproduce the main results using the shared data, users can:

• load the preprocessed datasets provided on Zenodo,
• bypass the raw data import steps in _targets.R,
• run the downstream targets corresponding to statistical models, tables, and figures.

This approach allows re-running the analyses while respecting data confidentiality constraints.

Repository structure

• _targets.R: full analysis pipeline, including raw data processing and downstream analyses
• R/helpers.R: general helper and utility functions
• R/data_preparation.R: data cleaning and dataset construction functions
• R/stat_models.R: statistical model definitions
• R/table_functions.R: table generation functions
• R/figure_functions.R: figure generation functions

Results

The main results are assembled in a Quarto document located in the results/ folder. The rendered HTML output is:

results/main_results.html

This file contains the main tables and figures of the study.

Author

Matthieu Rolland