Difference between revisions of "Study Data Validation and Submission Conformance"

From PHUSE Wiki
Jump to: navigation, search
m (Project Overview)
m (Project Leads)
 
(16 intermediate revisions by 3 users not shown)
Line 1: Line 1:
 
{{RightTOC}}
 
{{RightTOC}}
 
= Project Overview =
 
= Project Overview =
This project builds on previous successful modeling of the SDMT domains DM, VS, EX, TS, and AE and conversion instance data to Linked Data as Resource Description Framework (RDF). The original follow-on project was approved to extend the concept to additional domains. At the CSS 2019 conference, the team recognised that the project goals should change to focus on smaller, modular projects that can be released over time to demonstrate incremental value in a shorter time frame (instead of merely extending to additional SDTM domains). The proposed new focus increases cooperation with the Food and Drug Administration and the Maintenance and Support Services Organisation (MSSO, for MedDRA).  
+
This project replaces its predecessor "Going Translational With Linked Data (GoTWLD)", extending existing work in both clinical and preclinical areas by further developing the data models and instance data conversion. The project will include modeling FDA Technical Rejection Criteria to facilitate submission of data to the FDA. To view the project documentation page click [https://phuse-org.github.io/SENDConform/ here].  
  
This new proposal includes (but is not limited to) the following sub-projects:
+
Value Proposition<br>
 +
• Conformance errors for study data submissions to the FDA can be largely decreased using an ontology-based Linked Data model, including validation using Shapes Constraint Language (SHACL). <br>
 +
• The proof of concept will demonstrate creation of highly-compliant, high-quality preclinical and clinical study data for submissions, using a much more automated process than is currently available. Costs for data review, validation, and re-work will be greatly reduced. <br>
 +
• Separation of the results (instance) data from the standards data and metadata results in a version-free graph data structure for nonclinical studies and clinical trials results. CDISC-compliant data for submissions will be created by mapping the results data to the standards. Costs for recoding between CDISC versions will be drastically reduced.<br>
 +
• Metadata for submissions packages to the FDA can be standardised, validated, and semi-automated.<br>
 +
• The project provides a gateway for Knowledge Graph technology to support the FDA's Technology and Modernisation Action Plan (TMAP)<br>
 +
[https://www.go-fair.org/fair-principles/ F.A.I.R. Principles] will be followed with all work made available on GitHub, including a comprehensive website for documentation, explanation, and resources (https://phuse-org.github.io/SENDConform/  - currently under early construction, URL subject to change).  [https://protegewiki.stanford.edu/wiki/WebProtege Web Protégé] will be used for collaborative ontology development.
 +
<br>
 +
<br>
  
a) MedDRA as RDF  
+
=Project Dependencies=
Development and release of R and SAS scripts that convert MedDRA ASCII files to RDF. Due to licensing limitations, only the MedDRA RDF instance data that supports a subset of the test project (CDISCPILOT01) will be released online. The conversion scripts will be available publicly from a dedicated Github repository that will include documentation and usage instructions. The project team has the interest of and permission from the MSSO for this subject.  
+
• Participants with expertise in RDF creation and querying, including ontologies, SHACL, and related tools. <br>
 +
• CDISC SDTM and SEND terminology and domain knowledge.<br>
 +
• Participation from the FDA; specifically a sample or dummy eCTD file to support development of the prototype.<br>
  
b) Automating Conformance Checks for DM and TS domains in SEND and SDTM
+
Team members from the GoTWLD project will roll-over into the new project. With involvement from the FDA, and potentially academia, the project hopes to avoid the lack of staff resources which have hampered past initiatives.
The team will update existing SDTM ontology for the TS and DM domains to accommodate SEND. SPIN, SHEX, and SHACL will be evaluated for their potential to model conformance checks for both SEND and SDTM, working in cooperation with the Centre for Drug Evaluation and Research, with Dr. Lilliam Rosario as FDA sponsor. The project will evaluate the efficiency of replacing manual conformance checks using an automated Linked Data approach, with potential efficiencies for both FDA and industry.
 
 
 
c) Unique Identifiers for Pharma
 
A sub-group was formed after the PhUSE EUConnect18 to further develop the concept of "Study URI" presented by Kersin Forsberg and Daniel Goude from AstraZeneca. The group will develop methods and guidance for unique, universal, persistent identifiers for the pharmaceutical industry based on the RDF URI concept. Work is already underway (https://github.com/phuse-org/LinkedDataEducation/blob/master/doc/URIsForPharma.md) and a poster is planned for EUConnect19.
 
 
d) Study Ontology & Data Conversion
 
Work on Study ontology development may continue as time allows, with primary focus on deliverables for sub-projects a) , b) and c).
 
  
 
= Project Leads  =
 
= Project Leads  =
Line 22: Line 25:
 
|-
 
|-
 
|Tim Williams || Co-lead || UCB Biosciences || tim.williams@phuse.eu  
 
|Tim Williams || Co-lead || UCB Biosciences || tim.williams@phuse.eu  
|-
 
|Armando Oliva|| Co-lead || Semantica LLC || aoliva@semanticallc.com
 
 
|-
 
|-
 
|}
 
|}
Line 32: Line 33:
  
 
|Abdul Kadir|| Member|| Industry  
 
|Abdul Kadir|| Member|| Industry  
 +
|-
 +
|Aku Kallioniemi || Member || Aastat
 
|-
 
|-
 
|Amit Jain|| Member ||GSK  
 
|Amit Jain|| Member ||GSK  
 
|-
 
|-
 
|Andy Iverson|| Member|| Medtronic
 
|Andy Iverson|| Member|| Medtronic
 +
|-
 +
|Arpitha Hanumanthaiah ||Member|| Pointcross.com
 
|-
 
|-
 
|Dave Iberson-Hurst|| Member || Assero
 
|Dave Iberson-Hurst|| Member || Assero
 
|-
 
|-
|Lutz Weber|| Member ||Ontochem
+
|Dragomir Draganov || Member || Roche
 
|-
 
|-
 
|Erick Antezana||Member || Bayer
 
|Erick Antezana||Member || Bayer
Line 56: Line 61:
 
|-
 
|-
 
|Johannes Ulander|| Member|| S-cubed
 
|Johannes Ulander|| Member|| S-cubed
|-
 
|Mike Hamidi|| Member|| CDISC
 
 
|-
 
|-
 
|Katja Glass|| Member || Industry
 
|Katja Glass|| Member || Industry
Line 64: Line 67:
 
|-
 
|-
 
|Kurt Dauth|| Member||Boehringer-Ingelheim
 
|Kurt Dauth|| Member||Boehringer-Ingelheim
 +
|-
 +
|Lutz Weber|| Member ||Ontochem
 +
|-
 +
|Mike Hamidi|| Member|| CDISC
 
|-
 
|-
 
|Marc Andersen|| Member|| StatGroup
 
|Marc Andersen|| Member|| StatGroup
Line 70: Line 77:
 
|-
 
|-
 
|Naouel Karam|| Member|| Fraunhofer
 
|Naouel Karam|| Member|| Fraunhofer
 +
|-
 +
|Nicolas Dupuis || Member || Sanofi
 
|-
 
|-
 
|Nolan Nichols|| Member||Gene
 
|Nolan Nichols|| Member||Gene
 +
|-
 +
|Raphaël Noirfalise  || Member || JNJ
 
|-
 
|-
 
|Rashedul Hasan|| Member|| FDA
 
|Rashedul Hasan|| Member|| FDA
Line 102: Line 113:
 
|Objective || Timeline
 
|Objective || Timeline
 
|-
 
|-
|Project Initiation/Restart||February 2019/July 2019
+
|Supporting Ontologies|| February 2021
 
|-
 
|-
|MedDRA as RDF||PhUSE CSS 2020
+
|Technical Rejection Criteria Proof of Concept||February 2021
 
|-
 
|-
|Automating Conformance Checks ||PhUSE CSS 2020
+
|Submission Metadata Collection Proof of Concept ||February 2021
 
|-
 
|-
|Unique IDs for Pharma|| PhUSE CSS 2020
+
| Documentation and Resources (website)||February 2021
 
|-
 
|-
|Conclusion || PhUSE CSS 2020
+
|Project Conclusion || February 2021
 
|}
 
|}
  

Latest revision as of 10:07, 5 March 2020


Project Overview

This project replaces its predecessor "Going Translational With Linked Data (GoTWLD)", extending existing work in both clinical and preclinical areas by further developing the data models and instance data conversion. The project will include modeling FDA Technical Rejection Criteria to facilitate submission of data to the FDA. To view the project documentation page click here.

Value Proposition
• Conformance errors for study data submissions to the FDA can be largely decreased using an ontology-based Linked Data model, including validation using Shapes Constraint Language (SHACL).
• The proof of concept will demonstrate creation of highly-compliant, high-quality preclinical and clinical study data for submissions, using a much more automated process than is currently available. Costs for data review, validation, and re-work will be greatly reduced.
• Separation of the results (instance) data from the standards data and metadata results in a version-free graph data structure for nonclinical studies and clinical trials results. CDISC-compliant data for submissions will be created by mapping the results data to the standards. Costs for recoding between CDISC versions will be drastically reduced.
• Metadata for submissions packages to the FDA can be standardised, validated, and semi-automated.
• The project provides a gateway for Knowledge Graph technology to support the FDA's Technology and Modernisation Action Plan (TMAP)
F.A.I.R. Principles will be followed with all work made available on GitHub, including a comprehensive website for documentation, explanation, and resources (https://phuse-org.github.io/SENDConform/ - currently under early construction, URL subject to change). Web Protégé will be used for collaborative ontology development.

Project Dependencies

• Participants with expertise in RDF creation and querying, including ontologies, SHACL, and related tools.
• CDISC SDTM and SEND terminology and domain knowledge.
• Participation from the FDA; specifically a sample or dummy eCTD file to support development of the prototype.

Team members from the GoTWLD project will roll-over into the new project. With involvement from the FDA, and potentially academia, the project hopes to avoid the lack of staff resources which have hampered past initiatives.

Project Leads

Tim Williams Co-lead UCB Biosciences tim.williams@phuse.eu

Project Members

Abdul Kadir Member Industry
Aku Kallioniemi Member Aastat
Amit Jain Member GSK
Andy Iverson Member Medtronic
Arpitha Hanumanthaiah Member Pointcross.com
Dave Iberson-Hurst Member Assero
Dragomir Draganov Member Roche
Erick Antezana Member Bayer
Gaspare Mellino Member Roche
Giuseppe Di Monaco Member UCB
Hanming Tu Member Frontagelab
Ilaria Maresi Member TheHyve
Iraj Mohebalian Member Bayer
Jeremy Teoh Member Industry
Johannes Ulander Member S-cubed
Katja Glass Member Industry
Kerstin Forsberg Member Astrazeneca
Kurt Dauth Member Boehringer-Ingelheim
Lutz Weber Member Ontochem
Mike Hamidi Member CDISC
Marc Andersen Member StatGroup
Matthew Travell Member GSK
Naouel Karam Member Fraunhofer
Nicolas Dupuis Member Sanofi
Nolan Nichols Member Gene
Raphaël Noirfalise Member JNJ
Rashedul Hasan Member FDA
Siddharth Arthi Member Zifornd
Suhas Sanjee Member Merck.com
Sujit Khune Member Novonordisk
Susheel Arkala Member MMSHoldings
Todor Primov Member Ontotext.com
Tom Van der Spiegel Member JNJ
Vishnu Kollisetti Member PPDI

Project Updates

Provide project updates in this section.
Date: Description of Update

Objectives and Timelines

Objective Timeline
Supporting Ontologies February 2021
Technical Rejection Criteria Proof of Concept February 2021
Submission Metadata Collection Proof of Concept February 2021
Documentation and Resources (website) February 2021
Project Conclusion February 2021

Project Activities

This section can document project activities or serve as a jumping off point to other pages in the project.

Meeting Minutes


Archived Content