PASC Conditional Independence Testing

This package implements conditional independence testing procedures to identify codified concepts related to prior exposure to SARS-CoV-2 in 4CE Phase2.2 Data. Running this package necessitates refreshed Phase2.2 data up to Feb 1, 2022 inclusive of all patient cohorts.

4CE analysts at each participating healthcare system can follow the steps described bellow to run the analysis locally, using their healthcare system’s 2.2 data. This version of the package was not built for Docker. The summarized results are then shared and combined with those obtained from other healthcare systems. Those phenotype conditional testing results regroup PheCode, beta coefficient, standard errors, p-values, and prevalence estimates.

1. Install the package

Install the development version of Phase2.2PASCcondIRPackage from GitHub using remotes:

library(remotes)
remotes::install_github('covidclinical/Phase2.2PASCcondIRPackage', upgrade = FALSE)
library(Phase2.2PASCcondIRPackage)

2. Run the Analysis

Please note that the analysis can take up to two days to run, depending on your sample size.

dir.data = ""   # specify the input data directory, where the phase 2.2 data is saved
dir.repo = ""   # specify the output data directory, where the results will be saved
siteid = ""     # specify the site ID (capital letters)

# run the analysis
runAnalysis(dir.data, dir.repo, siteid, run.DML=T)

3. Submit the results

Your results will be saved under the path that you specify in dir.repo, as an folder titled "SITEID_conditional_testing_results" with .Rdata files inside. Please compress the folder into a zip format and share the results with @Harrison and @Clara-Lea via Slack.

Thank you very much for your contribution!