CVRG Eureka Analytics User Guide

From CVRG Wiki

Jump to: navigation, search

Welcome to the user guide for the CVRG Eureka Analytics, an online tool for extracting, transforming and loading clinical data into i2b2.

Contents

Quick Start Guide:

Overview

Welcome to a technology preview of the CVRG workflow application. The application provides an extract, transform and load mechanism that takes a clinical dataset conforming to a defined structure and loads it into an instance of i2b2. This VM is intended to demonstrate the functionality of the application and is not for production use. Future versions of the application are forthcoming that will add features needed to apply the workflow to your own clinical datasets.

This web application is hosted entirely on the virtual machine (VM) that you have installed and booted. The application has three tabs at the top of the screen, the Help tab, at which you are currently looking, the Workflow tab, which is the heart of the application, and the Administration tab, which is for troubleshooting (see below).

Running the application:

Navigate to the Workflow tab to run the application. The workflow starts with an Excel (xlsx only) spreadsheet and ends with data rendered into an i2b2 instance. There is a file list that is prepopulated with a spreadsheet (sample.xlsx) containing synthetic data. There is a 'PROTEMPA' box that represents the loading process, and an 'I2B2' box that represents the loaded i2b2 instance. Select the sample file by clicking on it, then click the play button between the 'PROTEMPA' and 'I2B2' boxes. This starts the workflow, transforming the sample.xlsx spreadsheet via PROTEMPA into i2b2. A red stop sign will appear in place of the play button for the duration of the loading process. Depending on the specifications of your computer, the loading process will range in duration from 10 minutes to 30 minutes (with the sample.xlsx). The loading is complete when the play button reappears. Click the 'i2b2' link in the I2B2 box to open its user interface. The i2b2 login screen is pre-populated with the correct username (i2b2) and password (demouser).

Data spreadsheet structure:

You may download the sample data spreadsheet to your computer to view the raw data by selecting the sample.xlsx file and clicking the download button (cloud with down arrow). The sample data spreadsheet represents data in a set of tabs for various broad classes of data (e.g., demographics, diagnosis codes, procedure codes, labs). The spreadsheet also contains a set of tabs that contain lists of codes representing the data types that are recognized by the workflow application. You are encouraged to browse these codes to get a sense of what data the application is capable of processing. These data types and data elements will appear in the i2b2 user interface as the data elements that are queryable. Note that the data in the sample file was synthetically generated and does not make sense clinically. Future versions of the workflow application will contain a detailed specification of the data spreadsheet file format.

Uploading data:

You may upload your own datasets that conform to the same structure as the provided sample spreadsheet and load them into i2b2. To test this, change some data in the sample (e.g., change the values of some diagnosis codes or lab test results) and then use the upload button (cloud with up arrow) to upload the spreadsheet back to the workflow application. You will have to rename the file before uploading. The workflow application will validate the data file and report any errors that are found. Correct any errors and upload the file again. Then run the workflow as described above, and look in i2b2 to see the changes. Using the workflow with your own data will be more fully supported in a future release. Note, especially if you are considering uploading real patient data to the application, that the application is currently provided without security, access control, firewalls, etc. You are STRONGLY encouraged to host the VM using a private, non-routable IP address on a desktop or laptop machine. A future version of the application will support appropriate security precautions.

Limitations:

The application currently has little in the way of security built-in as described above. It only supports one user at a time. In order to make the VM runnable on a wide range of computers, we have limited the default amount of RAM and disk space that is allocated to the VM. Instability can occur if your computer is hibernated while the VM is running, thus it is recommended that you shut down the VM (using VMWare Player, Workstation or Fusion's menu commands) beforehand.

Troubleshooting:

The Administration tab exists to provide some measure of control to restart components of the workflow that are not functioning normally. Behind the scenes, the workflow uses JBoss to host i2b2 and a web services interface to the workflow, and it uses Oracle-XE 10g to host your uploaded data once it is loaded into i2b2. The Administration tab contains two boxes for restarting the Webservice (includes i2b2) and Oracle instances. For each box, click the restart button in the lower right corner. Depending on the capabilities of your computer, restarting either service may take some time.

Feedback and bug reports:

As this is a technology preview, it is possible that something will go wrong. Please contact the developer team (richard.willard@emoryhealthcare.org or gsmilto@emory.edu) with feedback and issues.


Note: These instructions can also be found in the CVRG Eureka Analytics tool under Documentation.

Personal tools
Project Infrastructures