201705 Data Analytics

Return to September 2017 Proposals

Return to May 2017 Proposals

FHIR Data Analytics Track

Submitting WG/Project/Implementer Group

FHIR Management Group (FMG)

Justification

FHIR connectathons have been limited to developers to participate, or at least, for people who have tooled up development environments. But there are many people who come to HL7 who are interested in FHIR and need to learn more about it who are not in that position. This stream leverages the work done to prepare for the AMIA FHIR based datathon - uses the same set of tools to provide a data analytics platform that anyone who attends HL7 and is interested in informatics (most of us) can take part in.

The tools/platform provided, along with the test data, will allow participants to explore FHIR in practical ways.

This track can act as an introduction to the use of FHIR. The focus of the track content will be access clinical data to support informatics research activities and allow participants to evaluate how FHIR can support research. However, the tools and information that are included in this track provide a great place for beginners and non-developers alike to start using FHIR in meaningful ways. To help get participants started, we have provided some materials you can read to acquaint yourself with the topics and tools we will be covering in this track. This information, found below, will help you understand a little more about the tools and methods.

Links:

• AMIA announcement
• Corey's write up

Proposed Track Lead

Coordinator: Corey Spears . also see Zulip topic

Expected participants

• Health Intersections - terminology server
• - Synthetic data
• Chris Grenz - data analytics platform

Roles

In this track, users perform directed and undirected queries against the FHIR servers and data

Scenarios

These scenarios are just a starting point to stimulate your thinking at the outset, but participants are entirely free to explore others. The initial scenarios will explore searching for potential subjects for clinical trials who meet stated eligibility criteria (inclusion & exclusion). Example conditions to explore include asthma and dyslipidemia. Where you go beyond this will be up to you.

All codes supported by SyntheticMass can be found here https://goo.gl/VeQ9Bq

Scenario 1: Asthma Eligibility

Demographics

Age: 18-70 Gender: Both Race: All

Clinical Condition

Asthma History:

• Has the subject had sinusitis? Yes or No
• Has the subject had atopic dermatitis? Yes or No
• Has the subject had eczema? Yes or No
• Has the subject had seasonal allergic rhinitis? Yes or No
• Has the subject had bronchitis? Yes or No
• Has the subject had pneumonia? Yes or No

Diagnostic results of FEV1/FVC tests

Possible exclusion or inclusion:

• Current or Prior use of corticosteroids.
• Current or recent (last 2 years) use of an NSAID.
• Has had a nasal sinus endoscopy procedure performed.

Scenario 2: Dyslipidemia Eligibility

Demographics

Age: 18-70 Gender: Both Race: All

Exclusion criteria:

History of type 1 diabetes

• Other metabolic conditions such as obesity
• Pregnancy
• Other non-metabolic conditions, such as chronic kidney disease and systemic lupus erythematosus
• HIV Status

Lab Results:

Triglycerides > 200

Medications:

Medications that may cause dyslipidemia (Exclusion might be no prior use of these drugs within 6 months) Corticosteroids

TestScripts

N/A

Background, Tools & Methods

The Data

SyntheticMass – A “Realistic” Synthetic Patient Clinical Data Set SyntheticMass (https://syntheticmass.mitre.org) is an open-source, simulated Health Information Exchange (HIE) populated with one million realistic “synthetic residents” of Massachusetts. The data is free from cost, privacy, and security restrictions. It can be used without restriction for a variety of secondary uses in academia, research, industry, and government.

SyntheticMass provides:

• Realistic population health and demographic data at the city and town level
• The ability to drill down to individual synthetic patient data covering numerous common medical conditions
• APIs such as HL7 FHIR STU3 (https://syntheticmass.mitre.org/fhir) and HL7 CCDA

In SyntheticMass, you can use the dashboard (https://syntheticmass.mitre.org/dashboard) to drill down to individual synthetic patient data covering numerous common medical conditions. The FHIR Server you can use to query is at: https://syntheticmass.mitre.org/fhir (In the [clinFHIR.org clinFHIR.org]tool the server name is GoFHIR)

The conditions covered can be found here: [https://github.com/synthetichealth/synthea/wiki#currently-supported-diseases https://github.com/synthetichealth/synthea/wiki#currently-supported-diseases ] The list of all the concepts you may find in SyntheticMass can be found here: https://goo.gl/VeQ9Bq

The Tools

Below are a few tools we will be using for this track. They are provided to help you explore the data using FHIR.

clinFHIR – A Tool to Get Started on FHIR

clinFHIR is an educational tool that allows people to create or search for FHIR based resources and link them to tell a clinical story. clinFHIR is meant to help those not currently familiar with FHIR to start to understand what it is and how it can be used. clinFHIR is a front end tool that can be used to connect to any number of FHIR servers to explore and even store new data.

clinFHIR consists of a number of models that provide several capabilities that are beyond what we will be looking at for this track. For most of our work, the Query Tool (http://clinfhir.com/query.html) and Patient Viewer Tool (http://clinfhir.com/patientViewer.html) will be of most interest, however, you are welcome to explore the other modules (http://clinfhir.com). A description of those modules can be found here: (https://fhirblog.com/2017/02/20/new-ui-for-clinfhir/). Be sure to configure the FHIR server at the http://clinFHIR.org (SyntheticMass server name is GoFHIR).

Apache Drill – Direct Clinical BI on FHIR

This track introduces using the NoSQL to SQL engine Apache Drill to directly query FHIR formatted resources stored in NoSQL data stores. This track will:

• Introduce the technique: https://www.linkedin.com/pulse/mining-fhir-clinical-bi-chris-grenz
• Introduce practical considerations for accessing FHIR via Drill SQL
• Provide a demonstration environment loaded with MITRE’s SyntheticMass patient set

We will be using the same methodology and tools that were introduced in the recent AMIA Datathon more information about this approach can be found at the AMIA Datathon SQL Resources Site: http://amia.analystseim.com/

This site includes information about installing Drill related tools and the AMIA Datathon Drill environment.