Data Visiting Studygroup Meeting

Date: February 27, 2024 9 am EST

Meeting Chairs: Yann Joly  & Donrich Thaldar

Zoom meeting ID: 867 3160 3697

Zoom link: Zoom

Passcode: GA4GH

Zoom Recording: Zoom

Password: 3Vd*UAZ&

Attendees

1. Aida Falcon
2. Priya Ayyappaswamy
3. Anja Bedekar
4. Stephen Li
5. Shakuntala Baichoo
6. Hrefna Dögg Gunnarsdóttir
7. Alison Hall
8. Katherine Blizinsky
9. Diya Uberoi
10. Faith Kabata
11. Annalisa Landi
12. Donrich Thaldar
13. Owen Schaefer
14. Yann Joly
15. Tj Kasperbauer
16. Joe Flores-Toro
17. Robert Cook-Deegan
18. Melissa Darejeh
19. Shu ‘Sue’ Hui
20. Jessica Seegobin
21. Jordan Lerner Ellis
22. Adrian Thorogood
23. David Lloyd
24. John Clarkson
25. Claudia Gonzaga-Jauregui
26. Sonya Jooma
27. Brendan Behan
28. Palmira Granados
29. Reem Yunis
30. Richard Milne
31. Mutiat Afolabi
32. Anya Prince
33. Paul Ogendi
34. Francis P. Crowley
35. Cameron MacPherson
36. Peter Goodhand
37. Nathan Boyd
38. Ashley Hobb
39. Elena Ghanaim
40. Nunzia D’Ercole
41. Lindsay Smith
42. Anne Combon-Thomsen
43. Surakameth Mahasirimongkol
44. Emily Boja
45. Vasiliki Rahimzadeh

Welcome & Introductions

YJ welcomed the attendees and everyone introduced themselves in the chat.

Purpose of the Data Visiting group

YJ outlined the purpose of the Data Visiting studygroup. He highlighted that this group will work towards developing tools for data visiting. Data visiting however, is a concept that has yet to be captured in literature and is not well defined.  DT echoed the point that data visiting is a relatively new concept and presents different legal and ethical challenges compared to data download. While it offers the benefits of security and management, we still do not know if we have the necessary legal instruments to tackle data visiting and or even federated analysis.  He highlighted that there is a need to identify whether existing data regulations and legal infrastructure (e.g., data transfer agreements) adequately cover data visiting. By identifying this gap it will be possible to develop optimal legal and policy tools to support data visiting.

Discussion: YJ and DT then opened up the discussion. FC remarked that the Research Data Alliance has some guidelines on data visiting, and will share them when they are available. He explained that they are working on standard clauses as well. DT clarified that there shouldn’t be overlap with other groups, so an approach would need to be complementary to existing efforts.

YJ suggested that because data visiting is a new concept and that there aren’t many existing models, we may want to identify what approaches exist. We may want to consider Genomics England and All of US. Do they have a policy?

RM agreed with this point, but suggested that clarification around data visiting, secure data environment and even federated learning may be important. FC explained that there are differences amongst these concepts, even if there is overlap. It was thus believed that the group in its initial work, should also consider developing a glossary, to map the terms and identify the domains. For example, can Data visiting occur physically? These are questions that would need to be considered.

Countries also view data visiting differently and there are questions around responsibility. New data is being based on new data bases, so it becomes a question of border crossing again and then who has rights to the new data generated? These are additional concerns to be considered.

YJ and DT concurred that to tease out these issues, we may want to take a two pronged approach. It would be important to take case examples, ie. Genomics England and All of US to see how they function, how they approach and define Data visiting. Another good example would be the Japanese data from Riken being jointly analyzed by the BRCA exchange in the United States. It would however, be equally important to understand the needs of the community, so we should uncover why data visiting is even occurring.

Other concerns raised dealt with equity and the impacts that can occur from power imbalances and other underlying concerns. There may be overlap with other groups, which needs to be taken into account.

Apart from equity, it is also important to think about traditional and non-traditional concerns that can arise from data visiting and federated learning as they implicate matters of privacy and data control. A policy approach would be a novel contribution, but we would have to consider stewardship and how the existence of a data visiting model affects the way organizations steward data.

The group concluded with possible ways to go about this work.  These points included: literature review, understanding existing approaches (All of US & Genomics England) defining data visiting, glossary, webinars to bring in outside experts, differentiate between data access models and data visiting models.

Workshop Date & Time

DT suggested that the next meeting will be at Connect in Switzerland to clarify the scope of the work.

Key Takeaways:

• Sometimes it is not possible to share data, data visiting has thus emerged as a promising form of data sharing in genomics research.
• It allows the data recipient to access and analyze a provider's data in the provider’s online environment.
• This  use of algorithmic genomic data visitation however raises complex ethical and legal challenges that need to be identified and addressed.

Next Steps

• YJ and DT will reconvene to come forward with an initial plan.
• Background research will be conducted.
• The group will meet next in Ascona.