April 1, 2016 Meeting Notes

From PHUSE Wiki
Jump to: navigation, search

Present: Bob Friedman, Bill Varady, Mary Nilsson, Bill Houser, Kevin Snyder, Bob Dorsam, Chris Eley,

Notes from today’s discussion:

  • Script repository started in Google Code. Then was moved to GitHub.
    • Guidelines exist.
    • Quality status exists
  • Scriptathon experience
    • Good participation, but the end result is not developed in line with the big vision: flexible, re-usable code.
    • Their group is trying to decide if they should change their big-vision, or how can we reach the existing goal.
    • Can we partner with academia? Or other approaches.
    • Can we partner with TransCelerate?
    • It would be ideal if we could get a few people with dedicated time for a month or two.
    • Good to help people understand and use GitHub.
    • Mary would replace scriptathons with workshops focused on helping people learn to use GitHub
  • Bob Freedman’s experience with scriptathon
    • Folder called Qualified – only 1 that may be a non-real script.
    • Maybe 35 SAS scripts and 5 R scripts are in historical folder, but not much testing or further development progress.
    • Bob enjoyed the excitement of the work, he ended with a partly done script, spent the next several weeks finishing the script and submitting it. No one has looked at it since.
  • How do they identify scripting needs and prioritize them?
    • Came from white-paper team
    • They have 5 white papers
    • These describe a display, and helps to communicate the expectations, but it does not describe the requirements of what flexibility is needed.
    • These details are currently determined by the programmers.
  • Programming Guidelines exist on PhUSE
  • It is better to have a few people with more time compare do a lot of people with limited time.
  • How useful were the scripts from the FDA’s jump start?
    • They went into the repository and we can go look at the status.
    • They don’t necessarily match the goals specified in the white-papers.
    • It would be ideal to determine the requirements all together.
  • Challenges:
    • There is a knowledge hurdle to know how to use GitHub and the project’s structure and processes.
    • Requirements for scripts have not been specific enough
    • Script prioritization was based on people participating, not a broadly agreed process.
    • We need commitment from employers to volunteer dedicated staff for several weeks or a few months.
    • ADaM datasets might be better to use instead of clinical SDTM.
    • Goal of fully functional scripts (as opposed to starter scripts) may be to high.
  • Considerations for going forward:
    • Select a few (maybe 1) high-value scripts. Focus our effort on these.
    • If a script will be used by the FDA, the industry will want it.
    • Determine what quality of scripts we really want as our end.
    • Find an approach to get dedicated programmers.
      • Work with academia?
      • TransCelerate?
      • Pay with recognition since we don’t have money?
      • Understand what

Action Items

  • Kevin Snyder: Ask Hanming Tu (in email that Bob Dorsam sent) for a non-clinical folder, will try to put a script donated by FDA in it.
  • All: notify Mary if you are interested in the joint discussions with Clinical Script leadership team to find solutions to overcome the challenges.
  • Mary: Schedule joint time.