Repository files navigation

TheEDCimport package is designed to simplify the import and management of Electronic Data Capture (EDC) exports, particularly in clinical research settings. It is an opinionated framework, providing multiple streamlined tools for importing, cleaning, and checking your datasets.

# Install last version available on CRANinstall.packages("EDCimport")# Install development version on Githubpak::pak("DanChaltiel/EDCimport@v0.6.0.9016")

Use one ofread_all_sas(),read_all_xpt(),read_all_csv(), orread_trialmaster(), depending on the type of files in your export directory. You can then load your datasets into the global environment withload_database().

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

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

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

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 runfind_keyword("date") orfind_keyword("ecg").

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

lastnews_table() finds the last date of each subject throughout the whole database and inform onthe 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.

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

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

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

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