E-Rare Workshop on Data Sharing and Harmonization
On the 4th of May 2017 E-Rare – ERA-Net for research programmes on rare diseases, organized in collaboration with RD-Connect a workshop dedicated to data sharing and harmonization inviting researchers working on rare diseases to discuss on one hand the obstacles and on the other hand to accustom to several tools that are already available. The workshop was divided into three sessions.
Application of the data sharing policies and recommendations by the researchers
Initial discussions concerned General Data Protection Regulation and preparation of harmonized Code of Conduct for research. One of the most important aspects is the consent that participants in clinical research are asked to sign. The difficulty lies in the fact that ideally, a broad consent would let avoid re-consenting the patients. However, in reality the data should be collected for specified, explicit and legitimate purposes and thus patients should know exactly what they agree for.
Getting re-consent is a big obstacle and is always inefficient, but the recent publication of EURORDIS ‘You should at least ask’ suggests that patients want researchers to make efforts to re-consent whenever the purpose substantially changes.
A possible solution to avoid re-consenting is to have information provided on the Code of Conduct that describes potential circumstances under which the information/data will be used, such as “I don’t want to be involved in research using animals”; “I don’t want my data to be used in a commercial environment” etc.
The Code of Conduct also needs to address how to handle a patient’s potential desire to withdraw consent, how to handle the point at which a “child” becomes an “adult” (this varies by country), and how to handle data from deceased subjects. In addition, the Code of Conduct could state how far you can rely on an ethics committee to judge, rather than seeking re-consent each time.
Data sharing and harmonization tools and platforms
Online tools for sharing data were discussed with a particular emphasis on ways of querying data without sharing specific details about the participants. Profiles can be matched using phenotypic and genotypic information so that a researcher can find similar cases to their own at an abstract level. Information about where the specific details are held are provided so that the researcher can contact the data manager and request further details.
The problem with comparing data across isolated repositories is that computers cannot match data unless it is in a formal structure. Marco Roos and Rajaram Kaliyaperumal demonstrated a way of making data FAIR (Findable, Accessible, Interoperable and Reusable).
This can be achieved by using ontologies and coding that specify datasets, so that computers can classify and match them with datasets in other systems. In such a system, the data aren’t move (so they are not shared), but they becomes easier to query at the source.
Integrating, tools, platforms and patients participation: what is desirable and how to get there?
to be continued…