COVID-19 vaccine effectiveness against cardiac and thromboembolic complications following SARS-CoV-2 infection

Codes listed here are for the first staggered cohort study to assess the vaccine effectivness against Post Acute Covid-19 Sequelae. Notice that they assume that some cohorts are already instanciated cohorts. .JSON files are given as well as .csv with the cohort_definition_id for each.

Briew description

Scripts:

• AURUM_CDM_connection.R : These code is executed within the other scripts to stablish connection to the database and load some libraries.
• outcome_cohorts.R: in this script the outcome cohorts are created and instanciated to the results schema in the database.
• s01_dataAnalysis.R: this code does all the analysis once the data is obtained.
• s01_featureExtraction.R: this code extract covariates for the propensity scores and evaluated the balance with ASMD and NCO before and afetr the
  OW weighting.
• s01_forFeatureExtraction.R: this code is executed within the 's01_featureExtraction.R' script and contains libraries, functions and some variables to run the main script.
• s01_featureExtraction_AZ.R / s01_forFeatureExtraction_AZ.R: same but the vaccination cohort just includes AstraZeneca vaccine
• s01_featureExtraction_PF.R / s01_forFeatureExtraction_PF.R: same but the vaccination cohort just includes Pfizer vaccine

Folders:

• Data: contains 'NCO.csv' which is a list of condition occurrences for evaluating the Negative Control Outcomes, and 'estudi_pacs_03.csv' which contains the names of the table for each cohort_definition_id.
• BalancePlots: empty, it will be filled with plots created in the featureExtraction scripts.
• s01_RiskEstimates: contains multiple empty folders that will be filled with '.csv' files created in the script 's01_dataAnalysi' and contain the sHR, 95%CI, and SE.
• cohorts: cohorts as .JSON files.

Considerations

• The script 's01_forFeatureExtraction' saves the population weighted that will be used later for 's01_forFeatureExtraction_AZ' and 's01_forFeatureExtraction_PF'.
• I would recomend to instanciate the cohorts following the legend in 'estudi_pacs_03.csv'.
• Once the .JSON cohorts are instanciated, you can instanciate both population and outcome cohorts with the scripts for it.