EDCimport

EDCimport is a package designed to easily import data from EDC softwares in clinical research.

It is an opinionated framework, built to compute usual calculations given a database of several datasets sharing a common "Subject ID" key.

It can directly read from an export of TrialMaster.

Warning

This package is experimental and under active developement.
Retrocompatibility is not a priority for the moment.
For reproducibility, use renv to fix the package version.

Installation

# Install last version available on CRAN (once published)
install.packages("EDCimport")

# Install development version on Github
pak::pak("DanChaltiel/EDCimport")

Load the data

Your data usually comes as multiple files inside a specific directory.

Depending on the type of the files, you should use one of read_all_sas(), read_all_xpt(), read_all_csv(), or read_trialmaster().

For TrialMaster, you should request an export of type SAS Xport, with the checkbox "Include Codelists" ticked. This export should generate a .zip archive that you can use in read_trialmaster().

The resulting object db is an edc_database object containing all the datasets and some metadatas.

You can now use load_database() to import the list in the global environment and use your tables:

library(EDCimport)
db = read_all_sas("path/to/my/files/folder")
load_database(db) #this also removes `db` to save some RAM
mean(dataset1$column5)

Database management tools

EDCimport includes a set of useful tools that help with using the imported database. See References for the complete list.

Database summary

edc_lookup() returns a dataframe containing the number of rows, columns, patients, and the CRF name of each dataset.

Search the whole database

find_keyword() runs a global search of the database for a given keyword (or regex).

For instance, say you are looking for the "date of ECG" but don't know where it is, you can run find_keyword("date") or find_keyword("ecg").

It won't look into the actual data, though, as this would take too much computing power.

Get the last news date of each subject

lastnews_table() finds the last date of each subject throughout the whole database and inform on the date original dataset and column. It has arguments to avoid selecting irrelevant dates.

This is very useful to get the actual followup time when fitting survival analyses.

Data checking system

edc_data_warn() throws a warning if an inconsistency is found in a dataset. The interface allows to perform multiple checks and get a report as a CSV file.

Join helpers

As the primary key is almost always the Subject ID, join helpers were added to reduce code clutter. Currently, only edc_left_join(), edc_right_join(), and edc_full_join() are supported.

Shiny browser

edc_viewer() runs a shiny application that browses the whole database. The HTML interface is quicker and less cluttered than it would be in RStudio. It also allows filtering by Subject ID.

Swimmer Plot

edc_swimmerplot() creates a swimmer plot of all date variables of the whole database. This is very useful to find inconsistencies and outliers, especially with the plotly interactive output.