PDDI Info Model Task Force - Standard Sub-team minutes for 5/26

Previous Topic Next Topic
classic Classic list List threaded Threaded
1 message Options
Reply | Threaded
Open this post in threaded view

PDDI Info Model Task Force - Standard Sub-team minutes for 5/26

Richard Boyce-2
Hi Team,

Sorry for the delay....the minutes and link to meeting recording for the May 26th meeting of the standard sub-team of the PDDI Info Model Task Force are pasted below.

Kind regards, -R

Minutes for 5/26/2016 (Standard subgroup)

In Attendance:  Asiyah Yu Lin, Brian LeBaron, Oya Beyan, Jodi Schneider, Xia Jing, Michael Miller, Maria Herrero, Michel Dumontier, Oliver He, Daniela Oliveira, Ratnesh Sahay, Richard Boyce, Louisa Zhang, Elizabeth Garcia

Meeting recording:  http://goo.gl/Nv1Y6l


  • Introductions

  • Update from the Content sub-team meeting

    • Selecting drugs:  Goal is to choose interesting drugs that will be the focus of developing decision trees

    • Connecting clinical consequences, evidence, etc.

    • More than halfway done in terms of information categories

    • Information category discussion:  importance in identifying gaps in knowledge

      • Frequency of exposure, adverse events:  needs clarification

      • Mechanism not known:  discussion about relevance

    • Some decision trees have been created

    • Will also be getting engaged with stakeholder descriptions

  • Discussing progress on stakeholder descriptions

    • Drug Compendium Editor

      • Discussion:  ”editor” or “editorial process” is fitting, but multiple different roles are present

      • Diversity of potential user stories; e.g., clinical experts, medical librarians, people who do the publishing/design end, knowledge engineering and database and structuring end

    • Action Items

      • Instead of Drug Compendium Editor:  rename to  “Drug Information Editorial Board”

      • Add a user story about search from the perspective of a Librarian:  focusing on synthesis of information

      • Add a user story about Clinical Knowledge Engineer:  bridge between the clinical information and how that gets represented

  • User-Centered Definitions

    • “Clinical consequences” Qualtrics sent out

    • Comments under review; follow-up pending

  • Scope of Ontology Development and Related Efforts

    • Oya and Michael discussed minimum information model including MIAME and what has been done in other domains

      • MIAME and mimics are minimum information models; MAGE-ML

        • Minimal model of what’s required to be in the document, without specifying what needs to be in the model

        • 1 – what you must include (in English)

          • Using the MAGE object model (XML DTD), someone could include all of the information needed

        • 2 – providing a way of actually encoding and exchanging it

          • Java and PURL could take the information and represent it

    • W3C interest note as PDDI’s corollary to MIAME, describing the elements of the model

      • Introductory document with user stories, clinical descriptions

      • The clinical definitions may change as we go through the process

      • Creating deliverables – object model that makes it clear how to include the concept set of what to capture about drugs

        • W3C note as more human-readable description

      • Knowledge Representation Core Considerations:

        • Participation from sub-group (Matthias, Michal, Oya)

        • Organizing into a document form that will be how we build things moving forward

    • Balancing the need for vocabularies/ontologies with the use of free text

      • Model may be used by developers who develop clinical decision support systems; requirements for natural language processing if free text is used

      • Meetings so far have developed scope and definitions, evolving toward schema development

    • Computable components of info model

      • Essential elements = 8 categories, more or less

        • User-centered definitions, ontology definitions

        • Do not want to restrict the way that drug concepts and health info concepts are represented

        • Just need to code a drug (clinical drug level, or ingredient level, with formulation, rather than just at class level)

        • No current opportunity to structure alerts this way

  • Next Steps

    • Qualtrics survey:  Clinical roles stakeholder descriptions and user stories

      • Will be sent before Memorial Day weekend

      • 1-2 more surveys before the next Standards sub-team meeting

    • Discuss Asiyah’s email to the group

    • Knowledge Representation Core Considerations (Matthias, Oya, Michael)

      • Oya and Michael will work on edits

    • All team meeting

      • Use case:  Medication reconciliation (another team)

      • Will proceed to schedule and see if we can coordinate with them

Richard D Boyce, PhD
Assistant Professor of Biomedical Informatics
Faculty, Center for Pharmaceutical Policy and Prescribing 
Faculty, Geriatric Pharmaceutical Outcomes and Gero-Informatics Research and Training Program
University of Pittsburgh
[hidden email]
Office: 412-648-9219
Twitter: @bhaapgh