OHDSI Community Call 30Jul2019

MauraBeaton · July 26, 2019, 8:32pm

Hi Everyone,

Here’s the agenda for the tomorrow’s community call:

Introduction of OHDSI newcomers
Updates from the week
- OHDSI Network Research Studies: We’ll run through ongoing network research studies and ask study leads for the latest updates.
- 2019 OHDSI Symposium - Latest updates and announcements about the upcoming OHDSI symposium
- Looking for presenters for upcoming OHDSI calls
Presentation - @cce will tell us all about DataKnots 4

We are all geniuses up to the age of ten.
Aldous Huxley

UPCOMING COLLABORATION OPPORTUNITIES

Upcoming Deadlines
Rare Disease Clinical Outcome Assessment Consortium
https://grants.nih.gov/grants/guide/rfa-files/RFA-FD-19-029.html
Submission Due: July 29th

Drug Development Tools Research Grants
https://grants.nih.gov/grants/guide/rfa-files/RFA-FD-19-028.html
Submission Due: August 5th

2019 OHDSI Symposium - Titan Award Nominations
https://www.ohdsi.org/2019-titan-awards/
Nominations Due - Thursday, August 8th at 5pm

Dissemination of PCORI Improving Methods Projects to Advance Precision Medicine and Patient-centered Care - Proposal Due August 12th

Collaborating on a PCORI project to disseminate PCORI Improving Methods findings Researchers

My name is Yi Zhang. I’m a Senior Research Associate at MTPPI (a nonprofit research institute). I have been the PI of several PCORI-funded methods projects that aim to evaluate and apply causal inference methods for addressing comparative effectiveness research (CER) questions using observational data. Recently, PCORI approved my new LOI for the project – Dissemination of PCORI Improving Methods Projects to Advance Precision Medicine and Patient-centered Care – and invites us to send in a full…

Computer and Information Science and Engineering (CISE) Research Initiation Initiative (CRII)
https://www.nsf.gov/funding/pgm_summ.jsp?pims_id=504952&org=CISE&sel_org=CISE&from=fund
Proposal Due: August 14th

Upcoming Conferences
Machine Learning for Healthcare 2019 - August 8-10th in Ann Arbor, MI
https://www.mlforhc.org/

ASA Annual Meeting - August 10-13th in New York, NY
URL: https://www.asanet.org/annual-meeting-2019

2019 OHDSI Symposium - September 15-17th in North Bethesda, MD
https://www.ohdsi.org/events/2019-ohdsi-symposium/

HOW TO JOIN THE MEETING

Time: Every Tuesday from 12-1pm ET
Join the Webex: https://meetings.webex.com/collabs/#/meetings/detail?uuid=M59X2V1U61WC9ASID2Z5N3UT95-D1JL&rnd=96139.930901412523321531221112212141232121131213113112112121536
Join the call using VoIP
How to use the VoIP: http://wlc.webex.com/docs/job_aids/IntegratedVoIP_JA.pdf

OR

Join the call using the following teleconference details:
+1-415-655-0001 (US Toll)
Attendee Code: 199 982 907
Meeting details are also available on the wiki:
http://www.ohdsi.org/web/wiki/doku.php?id=projects:ohdsi_community

cce · July 26, 2019, 8:51pm

I’ll be presenting and asking for community feedback on three interrelated topics.

The creation of a regression test database using CDMv5, which is capable of returning hits for 3 cohorts discussed in the Book of OHDSI.

The conversion of 3 cohorts to use a high-level domain specific query language based upon DataKnots:

github.com

MechanicalRabbit/ohdsi-synpuf-demo/blob/master/cohorts/1770674.md

# Acute myocardial infarction events

This test case demonstrates the construction of a domain specific
language specific to the generation of OHDSI Cohorts. To get started,
we'll import some boilerplate definitions. This will create the
*DataKnot* called `sp10` that represents our test database.

    include("copybook.jl")

We implement the cohort definition as found in found in `1770674.sql`
that was produced by the JSON found in `1770674.json`, which defines a
cohort for persons experiencing acute myocardial infarction events.

## Concept Set Definitions

There are two concept sets involved, both of them SNOMED-CT codes for
*myocardial infarction*, one of them an older code (1755008) and then a
more current replacement (22298006). These codes are used to construct
a query, `is_myocardial_infarction` that will be `true` when a
`concept` matches one of these codes, or a descendent.

This file has been truncated. show original

github.com

MechanicalRabbit/ohdsi-synpuf-demo/blob/master/cohorts/1770675.md

# New users of ACE inhibitors as first-line monotherapy for hypertension

This test case demonstrates the construction of a domain specific
language specific to the generation of OHDSI Cohorts. To get started,
we'll import some boilerplate definitions. This will create the
*DataKnot* called `sp10` that represents our test database.

    include("copybook.jl")

We implement the cohort definition as found in found in `1770675.sql`
that was produced by the JSON found in `1770675.json`.

## Concept Set Definitions

This cohort defines 3 concept sets: (a) a diagnosis of hypertension,
(b) exposure to a hypertension drug, and (c) exposure to ace inhibitor.

    @define is_hypertensive =
         iscoded("SNOMED", 38341003)
    @define is_hypertension_drug =

This file has been truncated. show original

github.com

MechanicalRabbit/ohdsi-synpuf-demo/blob/master/cohorts/1770676.md

# New users of Thiazide diuretics as first-line monotherapy for hypertension

This test case demonstrates the construction of a domain specific
language specific to the generation of OHDSI Cohorts. To get started,
we'll import some boilerplate definitions. This will create the
*DataKnot* called `sp10` that represents our test database.

    include("copybook.jl")

We implement the cohort definition as found in found in `1770676.sql`
that was produced by the JSON found in `1770676.json`.

## Concept Set Definitions

This cohort defines 3 concept sets: (a) a diagnosis of hypertension,
(b) exposure to a hypertension drug, and (c) exposure to thiazide-like
diuretic.

    @define is_hypertensive =
         iscoded("SNOMED", 38341003)

This file has been truncated. show original

Plans of further work applying DataKnots to refine, evolve, document, and make performant this high-level query language, and integrating Cyclops so that queries could also include analytic aspects.

Thank you for the opportunity to discuss this fun topic and get community feedback. - Clark

cce · July 30, 2019, 3:59pm

The slides for today’s talk are at https://querycombinators.org/dist/ohdsi-dataknots-20190730.pdf

MauraBeaton · July 30, 2019, 9:21pm

Thanks everyone for joining today’s OHDSI call, and a big thanks to @cce for sharing his work with the group!

Clark’s slides have been uploaded here: https://www.ohdsi.org/resources/presentations/community-meeting-presentations/

And a recording of today’s call is also available here: https://drive.google.com/file/d/1l4fYkeAnrFObCET3vBL1V5r_D1sPwIf2/view?usp=sharing