README.md

In jreps/patientCluster: Performs patient clustering for patient cohorts using OHDSI CDM v5

patientCluster

Introduction

An R package for clustering cohort patients by medical history or creating data-driven medical topics (clusters of concept_ids corresponding to the same medical topic).

There is also a javascript D3 interactive visulisation available for viewing the results.

Features

• Aids the creation of topics (sets of similar concept ids)
• Takes a cohort as input and feature definitions (set of individual concept_ids or topics).
• Extracts the necessary data from a database in OMOP Common Data Model format.
• Performs kmeans, generalised low rank models or concensus clustering
• Includes functions for evaluating clusters and exporting into JSON format

Screenshots

Example Javascript Plot

Technology

patientCluster is an R package, with some functions implemented using h2o (http://h2o-release.s3.amazonaws.com/h2o/rel-lambert/5/docs-website/Ruser/Rinstall.html).

System Requirements

Requires R (version ? or higher). Installation on Windows requires RTools. Libraries used in patientCluster require Java.

Dependencies

• h2o
• DatabaseConnector
• SqlRender

Getting Started

1. On Windows, make sure RTools is installed.
2. The DatabaseConnector, h2o and SqlRender packages require Java. Java can be downloaded from http://www.java.com.
3. Install h2o as describe here: http://h2o-release.s3.amazonaws.com/h2o/rel-lambert/5/docs-website/Ruser/Rinstall.html
4. In R, use the following commands to download and install patientCluster:

```r install.packages("devtools") library(devtools) install_github("ohdsi/SqlRender") install_github("ohdsi/DatabaseConnector") install_github("ohdsi/patientCluster")

library("patientCluster") library("h2o") h2o.init(nthreads = -1, max_mem_size = '16g') ```

Getting Involved

• Vignette: Cluster examples
• Package manual: patientCluster.pdf
• Developer questions/comments/feedback: OHDSI Forum
• We use the GitHub issue tracker for all bugs/issues/enhancements

License

patientCluster is licensed under Apache License 2.0

Development

patientCluster is being developed in R Studio.

Beta

jreps/patientCluster documentation built on May 20, 2019, 10:46 a.m.