SS 01Jun2016

From PhUSE Wiki
Jump to: navigation, search


Meeting Record


When: 10:00 am - 11:00 am, 01Jun2016
Place: Teleconference
Facilitator: Hanming Tu
Scribe: Hanming Tu
Attendance: Hanming, Rebeka, Mary, Mat, Wendy, Peter, Jared, Gustav
Excused: Mike, Steve

Agenda
  1. Steering committee - Mary
  2. Review action items - All
  3. Review new project - Test data project
  4. Project Update - Project Leads
  5. Special topics
    1. Working Group Name Change

Name: Standard Analyses and Code Sharing

Vision: Leverage crowd-sourcing to improve the content and implementation of analyses for medical research, leading to better data interpretations and increased efficiency in the clinical drug development and review processes.

Goals: 1. Establish and maintain a publicly available repository for storing program code to be used as analytical tools for medical research. 2. Where gaps exist, develop recommendations for analyses and displays in areas that could benefit from crowd-sourcing. 3. Where gaps exist, develop code for recommended analyses and displays that could benefit from crowd-sourcing (to reside in the repository).

It is done!

General discussion

Steering committee update
  • Data visual analysis project may have some overlapped with whitepaper; get project description first and form the team later
  • Data visual analytic meeting on 5/17 will provide some clarity on the topic (Rebeka acting as liaison)
  • CDISC interchange: getting PhUSE particularly script WG more involved in CDISC Interchange - a session for each WG to give updates.
New Project - Test Data
  • Co-lead: Peter volunteered
  • Description from Peter's email:

Because I have expressed interest in the project about test data, I think I should clarify how I understand the project and how I would describe the objectives. I want to make sure my interest and what PhUSE and other folks were thinking is aligned. (Note this is not supposed to be a formal description but a summary of my thoughts about this.)

Testing and using appropriate test data are an essential part of any system development. Useful test data have two main objectives – they show that the system works correctly when the data are ‘good’ and they show that the system can handle data that are ‘not good’. (Whatever good and not good means in the context of the system.) Consequently, creating or acquiring test data should best be done with the specific system to be tested in mind. Several PhUSE projects describe procedures or even create components or subsystems that handle SDTM or ADaM datasets. As part of these efforts, a variety of test data are required. The typical fallback position of project team is to use data from the CDISC pilot project and/or anonymized study data provided by project team members. The PhUSE Test Data project proposes a different more systematic approach consisting of the following activities: 1. Define an initial scope of test data, i.e., which datasets are required (prioritized list) and what are the desired characteristics of these datasets (for example, incomplete, missing, or wrong data) 2. Create scripts (preferably R scripts) that will create the defined test datasets 3. Define an infrastructure that enables users to easily use the scripts to create test datasets 4. Define and implement a repository to store and share test datasets and to publish datasets and the appropriate metadata 5. Collect ‘real’ study data and publish in the repository While the first activities are clearly intended to apply a very systematic approach to creating test data with well defined features, the last one is adding a ‘real world’ perspective to the project – I think this should be lower priority but not neglected.

Any thoughts on this positioning of the project? Is this in line with what PhUSE/FDA/other people think? How else do you see the proposed project?

  • Action items
    • Polish the description - Peter
    • Submit the new project request - Peter
    • Send a link to Peter for submitting the project - Wendy
    • Find a replacement for Peter in P02 - Gustav: Nancy/Andrew/Eric Qi?
Project Update
  • P01 - SDA: Script discovery and acquisition: Rebeka
    • check the signup sheet to form the team
    • Sent out invite and wait for replies
    • First meeting will be in June in a week
    • Tasks:
      • 1) analyze the current index page and suggest a better way for organizing the scripts; https://www.ctspedia.org/do/view/CTSpedia/ClinAEGraph001
      • 2) conduct gap analysis between JS scripts and whitepaper: identify any issue and suggest any improvement
      • 3) conduct gap analysis between whitepaper and company's practice
    • Task: conduct GAP analysis between the JS scripts and whitepaper
  • P02 - RCD: Repository content and delivery: Gustav/Peter
    • Divide into two small teams
    • Had sub-team meetings and assigned targets to indivisuals
    • We have less than 10% with R skill; we decide to focus on SAS script first
    • Model of work: Use interns to review the docs and scripts? Peter to send the email to Wendy, Mary and Hanming;
    • Wendy to follow up. How to get work down is still open to discuss. Regarding use Interns - we could not use them based on the PhUSE board.
  • P03 - RGI: Repository governance and infrastructure
    • Had one meeting. Mike will give update next time
  • P07 - CPE: Communication, Promotion and Education
    • Jared, Wendy and Nina had a meeting
    • Created Google doc for upcoming events
    • Started working on paper abstracts on standard format
    • Coordinate the effort on creating abstracts, posters, etc.
    • Webinar for JS scripts
    • Paper for PhUSE annual conference has been accepted!
  • P08 - ADW: Analysis and Display White papers
    • QT whitepaper finished - it is final;
    • Working on AE whitepaper
    • AE whitepaper

Follow up

Action Items
  • FDA plan for releasing JumpStart scripts - Steve/Mat
    • Meeting every month and will bring it up again
    • We got 7 scripts. Are we getting more scripts from FDA?
  • Paper/Poster preparation:
    • PhUSE annual conference papers: Jared/Hanming
  • Plan for a webinar for JumpStart scripts in Fall 2016 - Wendy
  • Cross-check people in the signup sheet and Listbox groups - Mary/Wendy
  • WG name discussion - Mary will bring the names to be discussed
Next Meeting

Time: 10:00 am - 11:00 am (Wednesday)
We will conduct this meeting once per month.

Agenda for next meeting: Next meeting: TBD



Last revision by Htu,06/14/2016