Application of SEND Data for Analysis
Welcome to the site for the "Application of SEND Data for Analysis" project.
This project was conducted in 2015-2016 as part of the FDA/PhUSE Computational Sciences Nonclinical Working Group. The project is now closed, but the work done is available publically. This page provides the details and deliverables from the project.
To Learn more about the larger working group, and it's many other projects past and present, go to Nonclinical Topics Working Group.
The project focused on use cases for SEND data related to analyses provided on tables for general toxicology studies. Commonly, the nonclinical study analyses in industry practice are performed using raw data extracts, while the SEND datasets are created outside of this in a separate process. This project examined these purposes and identified some standard analyses that could be performed using the SEND package.
The intent was to understand how these analyses relate to SEND data, and demonstrate the data quality and fitness for use possible using SEND data sets for analysis. It was the project's objective that this effort can be used to create standard scripts and fitness for use checks on SEND data.
As the implementation and exchange of SEND data is becoming more prevalent in the industry, focus is directed towards the usability and fitness for use of the SEND data. The usability of SEND data is essential in order to harvest the benefits of more efficient review and analysis.
The project focused on individual SEND domains in a prioritized manor (FW, CL, MA, MI mainly). Scientific questions that may arise during review of SEND data were discussed. SEND domain specific analyses using example analysis keys (i.e. SEND variables) were identified to aid in getting the right answers from the SEND datasets.
Certain SEND variables lend themselves better to computational analysis by being subject to controlled terminology, formats and/or expectancy requirements by the standard, while other variables are under the control of the sponsor (either with population by need or choice or through sponsor controlled vocabularies). The project proposed recommendations for which variables are minimum information for any review of the data.
Analyses is defined for this project as review of SEND data, and does not refer to the statistical outcomes of the study.
To describe use cases for SEND data that may provide the industry with a common platform to communicate and assess how SEND supports data "fitness for use" and provides quality context for review of study data.
Transverse Collaborative Investigations
- Leverage similar deliverables from Phuse Standard Scripts work group (Development of Standard Scripts for Analysis and Programming). An example white paper from the Standard Scripts work group is linked here: SS P08 Demographics, Disposition, Medications White Paper
- Cooperation with the "Visualization of histopathology" project (link: Visualization of Group Related Differences in Histopatholgy Data)
- Find published clinical position papers within the content of dataset review and analysis (e.g. FDA position paper)
SEND Domains Examined
|Domain Name||Data Type||Source System||Link to document page|
|FW||Individual Non-rodent||PDS - ToxData||2016-08-12|
|CL (Qualitative Food)||Individual Non-rodent||PDS - ToxData||2016-09-09|
|MA||Group summary non-rodent||PDS - ToxData||2016-11-04|
|MA||Group summary non-rodent||PDS - ToxData||2016-11-04 continued|
SEND Variable Annotation Examples
|Document Name||Link to document page|
|Minute Template||File:Minute Template.docx|
|Tox tables mapped to SEND||File:Tox tables to SEND mapping.xlsx|
|Tox table illustration||File:Table illustration.docx|
|Sample BW report with annotation||File:Bw summary with individual.pdf|
|Sample BG Individual listing with annotation||File:IND BG Example.pdf|
|Sample LB Clin Chem Individual listing with annotation||File:LB Clinical Chemistry individual data.pdf|
|Sample LB urinalysis data - individual and summary||File:LB urinalysis examples reduced size.pdf|
|Sample CL data - individual and summary||File:CL examples2.pdf|
|Sample CL data 2 - Individual||File:Clin Obs Example.pdf|
|Sample CL data 3 - Individual||Media:Clinical_Sign_report.gif|
|Thomas' overview of study data||File:Framework for discussion of classification of information.pptx|
|Master Thesis, Thomas||File:Analysis of SEND LB data (MSc thesis).zip|
|Master Thesis, Thomas, Presentation||File:Analysis of SEND LB data (MSc thesis) - Examination slides final.pptx|
Work Group Participants
The team was established at the PhUSE CSS March 2015, composed of those interested in the scientific output and quality of SEND datasets, with the following areas of expertise:
- FDA reviewers
- IT vendors (with a special interest in the data quality aspect of SEND)
- Nonclinical data managers (with focus on the fitness for use aspect of SEND data)
- Biostatisticians (currently working or planning to work of SEND data)
Lead for this project was: Gitte Frausing (firstname.lastname@example.org).
- Gitte Frausing, Data Standards Decisions (lead), email@example.com
- Wenxian Wang, Xybion
- Dan Potenta, PDS Life Sciences
- Elyse Hoffmann, Boehringer-Ingelheim
- Jaga Virayah, Bayer
- Thomas Gade Bjerregaard, Novo Nordisk
- Rihab Kordane, Charles River Laboratories
- Rachel Harper, Covance
- Brian Argo, MPI Research
- Christy Kubin, MPI Research
- Paul Brown, FDA
- Anisa Scott, SAS Institute
- Kristi Johnson, PointCross Life Sciences
- Laura Kaufmann, PDS Life Sciences
- Robert Dorsam, FDA
|Date||Link to minutes|
Last revision by Laurenwhite, 2019-03-5