/covid19influenza

Primary LanguageROtherNOASSERTION

COVID-19 and Influenza Vaccination

This repository contains the scripts used to analyzed the data reported in the paper:

Influenza Vaccination and COVID-19 Mortality in USA

Claudio Zanettini*, Mohamed Omar*, Wikum Dinalankara, Eddie Luidy Imada, Elizabeth Colantuoni, Giovanni Parmigiani, and Luigi Marchionni.

Background

Analyses are performed using 4 set of variables :

  • knowledge-based: based on medical knowledge
    • set15: refered in the code as selected
    • set40: refered in the code as original
  • agnostic: selection based on variable importane estimated in random forest.
    • set12: refered in the code as rf_parsimonius
    • set38: refered in the code as rf_inclus

Files and folders:

code/scripts

Contains the scripts for retrieving, pre-processing and analyzing the data.

  • code/script/libraries_functions: loads the libraries (including covid19census).

  • code/script/functions_analysis: contains functions used to pre-process, analyze and summarize data. Functions are documented.

  • code/scripts/us_preprocess: this is used to retrieve and pre-process U.S data. Execution of this script returns the dataframe used for the analyses (dat_selected). A static copy of the dataframe is in data/all_raw.RDS.

  • code/strata_**set**.R: used to perform stratified analyses.

  • code/secondary_**set**.R: includes the analyses in which propesity score is divided in quintiles or tertiles and added to the linear model as a factor or left as a continous variable.

  • code/extra_analyses.R: 5-times 10-fold-cross validation resampling to identify most parsimonious model.

code/analyses.R

The file analyses.R executes all the scripts in code/scripts and print results tables.


Sources and data

The file data/all_raw.RDS contains a static copy of the data used for the analysis.

Details regarding the data sources as well as functions to extract updated COVID-19 data and aggregate them with other socio-economic and health related metrics can be found in the covid19census R package. Please refer to the package README or documentation for more information regarding the variables. The scripts used to import static data are reported in the package repository here.

** Note: use package renv and renv.lock file in parent folder to install the same versions of packages that we used.**