PseudoVet

VHA Innovation Project ID 407

Innovator/Architect: Will BC Collins IV Project Manager/COR: Brian Stevenson

*For current progress see proof_of_concept folder.

Introduction

PseudoVet is an automated patient data fabrication engine. It’s goal is to provide a set of active synthetic patients that can be used for healthcare software development and testing for applications that are geared towards VA’s VistA and Enterprise Heath Management Platform (eHMP) through the Veterans Health Administrations’ (VHA) Future Technology Laboratory (FTL) a publically accessible development environment. More information on the VHA Innovation Laboratory and FTL can be found here: http://vaftl.us

Background

Development against real patient data unnecessarily exposes patient health information (PHI) and personally identifiable information (PII) and cannot be used by developers outside of the VA network. Development against current fabricated data is not useful because the data sets are very old which require development teams to spend much time developing data sets to use in lieu of writing code. Typical fabrication of patient data is typically of little or no medical relevance. The development of a system that creates and updates synthetic patient data using a set of templates for various diagnosis would provide more relevant patient data for development that could be used both inside and outside of the VA network. Development outside of the VA network is desirable as it allows more collaboration with the Open Source community which is in-line with the VA’s Open Source Initiatives.

Why This Is A Good Idea

Every year there are a large number of development efforts to improve healthcare for our Veterans and more often than not, developers have to manually create test datasets to mimmick real patients while working on their projects. It takes an extraordinary amount of time for this to occur and the quality of the data is poor. There are various synthetic datasets that have been created but, require licensing and therefore not so easy to obtain and even when they are obtained they are not in a VistA system for teams to develop against them. By creating this project, there would be a substantial financial offset from having each development team generate data to test their projects against.

Technical Overview

PseudoVet’s fabricated patient records are created by random selection of diagnosis data such as service connected disabilities, symptoms, and thereby provides more clinically relevant fabricated progress notes, laboratory data, as well as surgical procedures, discharge, and other ancillary data. In addition to common clinical data related to specific diagnosis, PseudoVet also continuously schedules appointments, randomly no-shows patients, generate consults, means tests and other administrative activities that occur in a real patients record.

Architectural Overview

The PseudoVet system is to be comprised of the following components:

• Core Reference Database (CRD) – A database containing model and template data used for the automatic generation of synthetic patient data read more...
• PseudoVet Interface (PI) - A web based application and services to support the generation of synthetic patient and supporting data *It is hoped that the technology used for this effort is Node.js based.
• PseudoVet Database – A database where all generated synthetic patient resides
• Automation Services - Back-end services that generate and continuously update synthetic patient data
• Client EHR System Integration – Integration routines for data synchronization between the PseudoVet system and external Electronic Health Record Systems (EHR’s). The reference system to populate patient data into at this time is: OSEHRA VISTA using CentOS and Intersystems Caché database because it is what is used in production inside the Department of Veterans Affairs.

Building with Vagrant

PseudoVet can be build using the vagrant up command from the root of a cloned or downloaded repository. You must have Vagrant Installed to use this command.

• Rename Vagrant.sample to Vagrantfile

vagrant up

Deploy in AWS

• Edit awsconfig.sh
• set your access id, secret key, and all other parameters
• save your changes to awsconfig.sh
• next run the commands below

Note: If you don't want to use a parameter such as aws_elastic_ip, you will need to comment it out in the Vagrantfile

source awsconfig.sh
vagrant up --provider=aws

Once the build process completes, connect to the PseudoVet VM by issuing the following command:

vagrant ssh

Important Notes

• If you close your command prompt, terminal app, or shell, you will need to run 'source awsconfig.sh' again before you can execute 'vagrant ssh'
• If you reboot, via 'vagrant reload' or other means you will need to restart Cache' when the box comes back up:

  ccontrol start cache

The source will be located at /vagrant on the provisioned virtual machine

Building with Vortex (under development)

Vortex allows provisioning with VirtualBox as well as AWS. Since FTL is an AWS shop, it makes sense to provide a Vortex build. To use Vortex, install Node.js as well as Vortex by issuing the following command:

sudo npm install -g vortex

The default provider is VirtualBox.

vortex up

To build the system for AWS EC2, you will need to copy vortex.sample.json to vortex.json and edit the Vortex configuration file entering your AWS account data.

vortex --provider=Amazon boot

*More information on Vortex can be found here: https://github.com/websecurify/node-vortex/

Proof of Concept

See the proof_of_concept folder in the source for documentation on prerequisites, configuring, and running the scripts for that process.