19Mar2013 White Paper Breakout Session

From PhUSE Wiki
Jump to: navigation, search


When: 19Mar2013
Place: Silver Spring, MD
Facilitator: Mary Nilsson
Scribe: Mary Nilsson
Attendance: Wei Wang, Qi Jiang, Fabien Linay, Kirk Bateman, Sasha Ahrweiler, Jean-Marc Ferran (If I missed anyone, let me know)
Agenda:

  • Setting and confirming agenda - Mary
  • Final discussion on the first white paper – Mary
  • Defining the process of getting remaining white papers done and recruiting and assigning tasks to people - Mary

Anticipated discussion topics related to the first white paper (central tendency)

  • Should we include the change to min and change to max analyses for individual studies? Currently, it's only included as a recommendation for the ISS/SCS.
  • How should we position the change to min and change to max analyses? It appears we should position those as sensitivity analyses to the categorical analyses. If so, we might want to make sure they stay separate so they can be placed in appendices of reports and/or only included if "significant".
  • Potential refinements to the box plots: remove the reference lines (which is causing some confusion)? Instead, make the point red if it's outside of that subject's reference limits? Should we not change the width of the boxes with sample size (which is causing some confusion)?
  • ANCOVA versus other methods (e.g. MMRM)
  • Parametric versus non-parametric
  • When to use and not use transformations
  • Alignment on p-value section
  • Alignment on sections related to measurements after stopping med and discontinuation visit

Anticipated discussion topics related to next steps for first white paper

  • Communications around white paper 1, including posts, conferences, publication(s)
  • Request for scripts around white paper 1
  • How to get FDA agreement that if a company produced this set, they would find it acceptable

Anticipated discussion topics related to white papers 2-6

  • Finding leads for those that don't have them
  • Getting "active" teams for each white paper
  • Fine-tune white paper development process (how to request and receive feedback - new email code?, meeting frequency, how to get a broader set of reviewers)
  • Discuss target dates


Introduction

We introduced ourselves and discussed how we got involved.

General discussion


This summary is organized by the bullets provided in the agenda, not by order of how they were discussed.
Discussion around the first white paper

  • Should we include the change to min and change to max analyses for individual studies? Currently, it's only included as a recommendation for the ISS/SCS. [We will leave it as is unless we get additional comments to suggest otherwise.]
  • How should we position the change to min and change to max analyses? It appears we should position those as sensitivity analyses to the categorical analyses. If so, we might want to make sure they stay separate so they can be placed in appendices of reports and/or only included if "significant". [We will position them as sensitivity analyses to the threshold analysis - and explain why]
  • Potential refinements to the box plots: remove the reference lines (which is causing some confusion)? Instead, make the point red if it's outside of that subject's reference limits? Should we not change the width of the boxes with sample size (which is causing some confusion)? [We will remove the reference lines and make red points for out-of-range values; we can say reference range can be added if there's only one; we won't change the width of the boxes since it's confusing]
  • ANCOVA versus other methods (e.g. MMRM) [Did not discuss other than saying it would be nice to get FDA feedback]
  • Parametric versus non-parametric [Did not discuss]
  • When to use and not use transformations [Did not discuss]
  • Alignment on p-value section [Briefly summarized and there was no disagreement as to how this was handled in the white paper]
  • Alignment on sections related to measurements after stopping med and discontinuation visit [We discussed this when Chuck Cooper joined our group. We will make some modifications to the section to expand more on points to consider - Chuck gave an example of liver enzymes increasing after drug stopped - the study was for an antibiotic which was a very short study. Points to consider - length of study, likelihood of going on another therapy, half-life of drug]
  • Other topics discussed: what to do if there are too many visits for the boxplot? Do we need to specify what we recommend for the report versus safety review purposes? Need to be clear that these only include subjects with a postbaseline result. We might want to put "Analysis" in the title. Chuck Cooper though it was appropriate that the safety population was those that received at least one dose.

Discussion topics related to next steps for first white paper

  • Communications around white paper 1, including posts, conferences, publication(s) [We agreed we should publish. It sounds like FDA may not be able to be identified as a specific author, but can be included when the author is a Working Group. We agreed we need to aggressively get the word out at conferences. We will continue this topic moving forward. We should definitely include the PhUSE conference in October - PhUSE does have a statistics stream]
  • Request for scripts around white paper 1 [The other projects within our Working Group will help author scripts. They will network with Wei Wang to get the code that was used for the white paper.]
  • How to get FDA agreement that if a company produced this set, they would find it acceptable [Chuck Cooper agreed to review the white paper and will think of others to forward to. It is recognized that there is a lot of variability within the FDA. The safety group at FDA might be interested.]

Discussion topics related to white papers 2-6

  • Finding leads for those that don't have them [Qi mentioned that Seta from Amgen might be able to lead the AE white paper. We will have to recruit others for the other papers that don't have a lead yet.]
  • Getting "active" teams for each white paper [We discussed the need for networking to get additional active members.]
  • Fine-tune white paper development process (how to request and receive feedback - new email code?, meeting frequency, how to get a broader set of reviewers) [We discussed that it might be best for the lead to write the first draft and then recruit help to fill in the holes. That seemed to work for the first white paper - when someone was asked to help with a particular section, they were willing to do that. We still need to work on the process for receiving feedback on white papers. Perhaps leverage QSPI. Use PhUSE membership list for reviewing? Maybe find out who authored the safety reviewer guidance - perhaps they will want to ge involved. Create a LinkedIn Group?]
  • Discuss target dates [No specific discussion other than we're hoping to have all remaining white papers final or near final before the next conference.]
  • Fabien went over the progress to date on White Paper 6 - the PK white paper. He and Ingrid have made a fair amount of progress. It's not ready yet to be considered a first draft.

Additional topics

  • Address special topic analyses either as v2 to the white paper or its own white paper (e.g., next step when you find a signal)
  • Create a time to event white paper?
  • Address number of occurrences for events somewhere
  • Got on a tangent on Program Safety Analysis Plans - maybe publish some of the PSAP success stories. Maybe FDA can help create real examples of when a PSAP could have prevented a problem.

Follow up

Allow additional time for FDA (Chuck and possibly others) to review the first white paper. Finalize first white paper (Mary) - We will stay with the May target date. Mary will recruit help where needed. Our project will work with the communication project to develop an aggressive communication plan. In the mean time, all of us should think about presenting at any conference we plan to attend. Fabien will continue drafting the PK white paper. He will let us know when it is ready for additional reviewers. All of us to help him with the networking to get reviewers. Mary will set-up project meetings.


Last revision by MaryNilsson,03/22/2013