Efforts like GA4GH enable the sharing and reuse of human medical and genomic data, which can greatly accelerate health research. However, the purposes and protocols applied to reused data usually differ from the purposes and protocols for which the data was originally collected. Since these differences are often material for the research participant, it is necessary to factor in participant consent choices when authorizing data reuse.
The platform presented here, DyCons (Dynamic Consent), enables dynamic consent on distributed health data infrastructures, which facilitate the sharing of human data between institutions. Dynamic consent systems present both technical and procedural challenges, particularly in the multijurisdictional Canadian context. For platforms such as CanDIG (Canadian Distributed Infrastructure for Genomics), it is necessary to develop technologies and processes for compliance with heterogeneous institutional and provincial requirements, as well as for compliance with the consent choices of individual research participants. To factor consent into the processes of computer-mediated data-sharing, we consider the different roles that human actors play in consent-solicitation, data-use authorization, data stewardship, and querying.
DyCons’ design minimizes the data visible to each type of actor, reducing the potential for system-wide data leaks. Data-access authorization depends upon context collected from multiple sources, including the prospective data user’s institutional identity claims, the Data Access Committee’s dataset-level access permissions, and the participant or clinician’s individual-level access permissions. DyCons is able to support models of both informed consent and broad consent, while keeping a research datum in the context of its consent metadata even at query-time.
Through these design principles, DyCons is able to extend a distributed data-sharing infrastructure to take into account participant rights to autonomy, while still being able to adapt to the differing data governance needs of participating institutions. By improving participant trust in these systems, DyCons can facilitate the broader sharing of health data.