PhUSE Nonclinical WG – Application of SEND data for analysis

AGEnDA/Minutes

/
2015-04-24
/ Telecon/WebWebEx / 10:00 am – 11:00 am (EST)
4:00 pm – 5:00 pm (CET)
Meeting Organized by / Gitte Frausing & Laura Kaufman
MEETING PURPOSE / Discussion of use cases for analysis of SEND data
Attendees / Present / Name / Company
X / Anisa Scott / SAS Institute
X / Dan Potenta / Novartis
X / Gitte Frausing / Data Standards Decisions
X / Jillian Sanford / PointCross Life Sciences
X / Laura Kaufman / PDS Life Sciences
Lynda Sands / GlaxoSmithKline
Montserrat Cases / Bayer Pharma
X / Paul Brown / FDA
Rachel Harper / Covance
Richard Buchanan / PDS Life Sciences
X / Rihab Kordane / Charles River Laboratories
Robert Dorsam / FDA
X / Thomas Gade Bjerregaard / Novo Nordisk
X / Wenxian Wang / Xybion

Agenda topics

1.Prioritization of SEND domains

/

We decided to start with BW – Body Weights and get an approach for the remaining domains. When finalizing BW, we will revisit the other SEND domains for subsequent prioritization.

2.How to approach the analytics with a SEND perspective

/

Home assignment for next time: Find a representative ways of visualizing Body Weights data and annotate it with SEND variables.

We discussed several ways to approach getting the SEND analytical keys (i.e. variables) listed.

It was suggested to use an analytical tool (like Spotfire or JReview), create a table and assess its usefulness.
Another approach was to annotate known ways of visualizing the data, like the individual and summary tables in our respective nonclinical study reports.
An example of such an annotation is shown below:

It was argued that all information in a table or figure may not be captured within just one variable or indeed in the SEND datasets at all.
This is something to be mindful of when conducting this exercise. The exercise of annotating a table from a report is to facilitate the generation of “compiled SEND tables” for review. When analyzing SEND data, the purpose is not necessarily to create an exact replicate of the table in a report, but to get data in a SEND dataset to a reviewable state.
Some things to consider:
  • Can the table that you are attempting to annotate be represented a bit differently, so it fits to SEND variables and still be appropriate for review of the data (example: using VISITDY instead of manual timing labels “Week 1”, “Week 2”…)?
  • Is informationmissing in the SEND datasets that is necessary for compiling the data correctly? I.e. is the data that must be grouped together somehow identified in the dataset package (could be through Trial Sets and Sponsor Defined Treatment groups)?
  • Is information missing in the SEND datasets that is necessary for visualizing the data correctly? I. e. are appropriate labels, designators or flags missing from the SEND data?
Everyone are free to upload the outcome of the BW annotation on the Wiki.
Please add another row in the table under “Working Documents” and list “your name” – BW annotation.
If you are not allowed to share the page from a report, you can just present the outcome of your analysis (which SEND variables that you identified) at the next TC.

Page 1 of 2