README.md

In mi-erasmusmc/patientLevelPredictionUI: Package for a UI for patient level prediction using data in the OMOP Common Data Model

PatientLevelPredictionUI

Introduction

An UI built on top of the R package for building patient level predictive models using data in Common Data Model format.

Features

• Takes a cohort and outcome of interest as input.
• Extracts the necessary data from a database in OMOP Common Data Model format.
• Uses a large set of covariates including for example all drugs, diagnoses, procedures, as well as age, comorbidity indexes, etc.
• Large scale regularized regression to fit the predictive models.
• Includes function for evaluating predictive models
• Supported outcome models are logistic, Poisson, and survival (time to event). TO ADD.

Technology

PatientLevelPredictionUI is a Shiny application running on R.

System Requirements

Requires R (version 3.2.2 or higher). Installation on Windows requires RTools. Libraries used in PatientLevelPrediction require Java.

Dependencies

• PatientLevelPrediction
• Cyclops
• DatabaseConnector
• SqlRender

Getting Started

1. On Windows, make sure RTools is installed.
2. The DatabaseConnector and SqlRender packages require Java. Java can be downloaded from http://www.java.com.
3. In R, use the following commands to download and install PatientLevelPredictionUI:

r install.packages("shiny") install.packages("devtools") library(devtools) install_github("ohdsi/SqlRender") install_github("ohdsi/DatabaseConnector") install_github("ohdsi/Cyclops") install_github("ohdsi/PatientLevelPrediction")

Getting Involved

To Add

License

PatientLevelPredictionUI is licensed under Apache License 2.0

Development

PatientLevelPredictionUI is being developed in R Studio.

Alpha

Acknowledgements

• To Add

mi-erasmusmc/patientLevelPredictionUI documentation built on May 22, 2019, 8:57 p.m.