We’ve just wrapped up another successful CDISC + TMF EU Interchange event! This year’s event, which was attended by 352 people, featured parallel sessions dedicated to the Trial Master File (TMF) Reference Model, which has been a part of CDISC standard development initiatives for over a year now. This addition is crucial, as documentation is a serious and important aspect of our industry. Each of us must contribute to this documentation effort and stick to the principle “if it is not documented, it did not happen.”
Though not particularly my area of expertise, I attended one of the TMF parallel streams, specifically the “TMF Essentials” stream. Here, I gained insights into the TMF standard and realized the urgent need for standardization in how TMFs are managed and structured. It appears that companies still employ various and different approaches. While we at Cytel, as a biometrics CRO, do not directly manage TMFs, we generate artifacts that eventually find their place in the sponsor TMF. Hence, it’s crucial to be aware of content requirements, particularly what constitutes “Essential Documents,” and to proactively engage with our sponsors in TMF planning at the trial’s outset.
For biometrics colleagues in the US, there will be specific training on the “TMF Reference Model for Biometrics Professionals‚” at the next CDISC US Interchange in October.
This year, Cytel once again played a significant role in organizing the event. My colleague, Silvia Faini, and I are part of the CDISC EU committee responsible for defining the agenda. In addition to moderating the “ADaM” and “Piloting JSON” streams, we also delivered two presentations:
- “10 Years of FDA Study Data Technical Conformance Guide. Have You Missed Any FDA Data Submission Requirements?” (For further details, refer to my test presentation)
- “Visualizing ADaM,” where we showed our ADaM implementation approach (more details will follow in a future blog post)
Cytel was also among the exhibitors, with my colleagues Russel Brown and Cyril Sombrin also participating.
The event commenced with visionary presentations by the new CDISC president, Chris Decker, and a CDISC roadmap presentation by Peter Van Reusel, Chief Standards Officer at CDISC. From Chris’s presentation I’d like to highlight his possible “back to the future” comment:
“Provide regulatory agencies a connected and traceable submission package allowing reviewers to click their way from the clinical study report to analysis to data to protocol increasing confidence in the data and results.”
The keynote speech by Rob DiCicco from TransCelerate BioPharma Inc. also set the tone for the event “Our Industry Journey towards Digital Transformation.”
I then jumped into the “Navigating Submission Pathways” stream as I had scheduled my presentation. Worth mentioning from this stream was the presentation from Jintao Shi “Understand PFDD Guidance Series and the Impact on Data Submission,” which discussed two recent FDA guidance documents and their implications on SDTM and ADaM datasets packages:
- Submitting PRO Data in Cancer Clinical Trials
- Submitting Clinical Trial Datasets & Documentation for COA Using IRT
While I was engaged in the “Navigating Submission Pathways” stream, Silvia attended the CDISC CORE Workshop, focusing on the CDISC Open Rules Engine project. She found the workshop highly impressive and is thinking of joining the team. Many companies are already integrating CORE into their processes and tools, as discussed in the “Core Implementation” stream. Novo Nordisk, SGS, Argenx, and OCS shared their experiences as the first implementers of CORE rules. Clear rules are urgently needed to provide precise outcomes and improve reporting.
Silvia moderated the “Pivoting JSON” stream. There are high expectations from this project within our industry, as there is a pressing need for a better method of exchanging datasets beyond SAS XPT. CDISC, with Sam Hume, discussed the results of the Datasets JSON Pilot, an initiative that actively involved the US FDA. While a full “readout” will be disclosed at the PHUSE CSS conference in June 2024, Sam shared the progress made so far. The pilot demonstrated that Dataset-JSON can transport information successfully (verified by FDA testing on their environment). However, improvements are needed in guidance and the standard itself, along with enhancing tools such as Python conversion software. The team is still seeking volunteers to join version 1.1 of the pilot.
I also attended “Submission Experience” stream where my Italian friends from Chiesi Farmaceutici, Rebecca Moretti and Rocco Calabro’, presented their experience submitting data packages to the NMPA Chinese health authority. Their presentation was recognized as the best of the entire conference. Another notable presentation came from Marilina Castellano of the Uppsala Monitoring Centre, who provided details about WHODrug and its implications for SDTM mapping.
On the regulatory side, all three main health authorities were represented live. Though FDA and PMDA provided no major updates, it’s worth noting that PMDA mentioned “Form A” is no longer required for submission. EMA provided updates on current initiatives towards data standardization and submission, including the “Submission of individual patient data from clinical trials.” Eftychia Eirini Psarelli presented some key learnings from the ongoing pilot project and announced the interim report’s release by Q2 2024. Final decisions regarding EMA requirements for dataset submission are expected by late 2024 or early 2025.
From the “ADaM” stream that I moderated, I’d like to highlight Cedric Davister’s (Merck KGaA) and Alexandra Cochinaire’s (Johnson and Johnson) presentation, both members of the PHUSE Estimands Working Group. They provided an overview of the draft white paper on the Implementation of ICH E9(R1) Estimands Framework using Data Standards. Their focus was on the implications of the Estimands Framework on ADaM datasets and, in particular, how to handle “Intercurrent Events.”
Many interesting presentations were scheduled on the agenda that I was unable to attend. Notably, the entire “AI in Standards” stream explored the use of Large Language Models (LLM) in aiding standard implementation, presented by SAS in collaboration with Deda Group and similarly with the Novo Nordisk with a practical application of AI agents as Co-Scientists.
I also had the pleasure of coordinating the poster session, reviewing drafts, and suggesting improvements to the authors. We had a total of 16 posters covering topics ranging from the use of CDISC Library to SDTM and ADaM implementation showcases, as well as Open-Source topics. The winner was Siddhant Prabhu from Novo Nordisk with “Data Quality Considerations when Providing Large Volumes of Consolidated Clinical Trial Data for Secondary Use.” His work provided an interesting application to assess quality and adherence to SDTM standards in current and historical Novo Nordisk studies.
CDISC members can find copy of the slides presented on the CDISC website.
Lastly, while details are yet to be finalized, it was announced in the closing session that the next CDISC + TMF EU Interchange will take place in Geneva, just 800 meters from the Cytel Geneva office. I look forward to seeing many of you there next year on May 14–15!
I would like to extend my gratitude to my colleague, Silvia Faini, for providing insights into the presentations she attended.
Interested in learning more?
Download Angelo Tinazzi’s new ebook, “The Good Data Submission Doctor on Data Submission and Data Integration to the FDA”: