EHRtemporalVariability

EHRtemporalVariability is an R package for delineating temporal dataset shifts in Eletronic Health Records

What is this repository for?

The EHRtemporalVariability package contains functions to delineate temporal dataset shifts in Electronic Health Records through the projection and visualization of dissimilarities among data temporal batches. This is done through the estimation of data statistical distributions over time and their projection in non-parametric statistical manifolds, uncovering the patterns of the data latent temporal variability. Dataset shifts can be explored and identified through visual analytics formats such as Data Temporal heatmaps and Information Geometric Temporal (IGT) plots 1-3. An additional EHRtemporalVariability Shiny app can be used to load and explore the package results and even to allow the use of these functions to those users non-experienced in R coding.

Package’ Status

Copyright: 2019 - Biomedical Data Science Lab, Universitat Politècnica de València, Spain (UPV) - Department of Biomedical Informatics, Harvard Medical School (HMS)

Documentation

Citation

A paper describing the EHRtemporalVariability package has been submitted. If you use EHRtemporalVariability, please cite:

Sáez C, Gutiérrez-Sacristán A, Isaac Kohane, García-Gómez JM, Avillach P. EHRtemporalVariability: delineating temporal dataset shifts in electronic health records. (Submitted)

The original methods and case studies using the approach are described here:

Download

Install the latest released version from CRAN

install.packages("EHRtemporalVariability")

Download the latest development code of EHRtemporalVariability from GitHub using devtools with

devtools::install_github("hms-dbmi/EHRtemporalVariability")