By Andreas Bruns and Eva C Winkler.

Informed consent is a central principle of medical research ethics. Traditionally, consent is required to respect the autonomy of human research subjects—their right to make their own, informed decision about whether or not to participate in medical research. However, this principle has come under significant pressure with the rise of biobanks and data infrastructures.

Biobank and data infrastructure research (from here on, ‘B&D research’) differs from traditional forms of scientific research involving human subjects. Instead of conducting research directly on the subject, B&D research uses biosamples (such as blood, tissue, or cells) and related data (such as digitalised health information) from large groups of subjects to perform comparative analyses and understand diseases and health conditions on a larger scale.

Currently, many global efforts aim to unlock the potential of B&D research, particularly the potential of medical Big Data to improve healthcare. But one central question continues to haunt these efforts: How should researchers go about seeking agreement from subjects for the ongoing and potentially open-ended research use of their samples and health data? They cannot, without seriously obstructing B&D research, seek informed consent for each research project separately.

Different models of consent have been proposed to address this question. In many contexts, researchers rely on ‘broad’ consent, where subjects agree to the storage and ongoing use of their samples and health data without being asked to consent to each specific use. Subjects may withdraw their consent at any time, and researchers are responsible for ensuring that all research falls within the limits of the kinds of uses that were previously agreed upon.

While broad consent facilitates research, some critics argue that it would not truly respect autonomy. Subjects do not know, nor can they control, which studies will use their samples or data. How, then, can such broad consent offer the levels of information and control necessary for autonomous decisions?

In response to this criticism, alternative consent models, such as ‘dynamic’ consent, have emerged. In the dynamic consent model, subjects first agree to have their samples or data included in some type of study platform; subsequently, they receive information about new studies and can decide—study-by-study—whether their samples or data may be used. In other words, instead of broadly consenting to B&D research, subjects get to personalise their consent, deciding for themselves how they want to engage with medical research.

Concerns have been raised about the practical challenges to implementing dynamic consent. Additional concerns focus on potentially undesirable consequences, such as dynamic consent becoming a burden on both researchers and research subjects, or that reliance on digital consent interfaces might alienate those less accustomed to digital technologies. Yet, dynamic consent has found many supporters among lawyers, ethicists, and policymakers who believe it represents a major step in the right direction.

Our paper wants to provide a useful analysis and critical assessment of what seems to be the underlying assumption of much of the support for dynamic consent: that dynamic consent is preferable to broad consent because it better respects the autonomy of research subjects.

There are broadly two concerns with this assumption, depending on how we understand the supposed problem with broad consent. First, the problem might be that broad consent cannot accommodate people’s different informational needs. While some studies suggest that many people would be happy to provide their broad consent to research, some might have reservations and wish to know which specific studies are covered by their consent. In this case, however, there is no reason to think that consent needs to be dynamic, just that information should be provided in an ongoing manner.

Second, the problem might be that broad consent does not offer subjects enough control over how their samples or data are used. People need to be able to specify which individual studies they want to contribute to; a broad ‘yes’ or ‘no’ to collective research activities is not enough. However, while dynamic consent would increase the occasions on which subjects can make consent decisions, it is neither obvious that dynamic consent would introduce options for choice that meaningfully enhance subjects’ autonomy, nor that it is morally appropriate to act on every study-specific consent choice. This is so because autonomous choices are a certain species of choices but dynamic consent treats autonomous choices on a par with choices that are random, ill-considered, arbitrary, or even morally objectionable.

The paper suggests that dynamic consent might not be a step in the right direction after all. The case of B&D research shows that it is time to reconsider the role of informed consent in state-of-the-art medical research. Reconsidering this, we should also be open to discussing alternatives to the traditional opt-in consent for B&D research. One such alternative is the opt-out model proposed for the currently planned European Health Data Space (EHDS). Unlike an opt-in model that requires explicit consent, an opt-out model allows samples and data to be used for certain kinds of research unless subjects have explicitly refused. Strong safety and control measures, well-curated information sites, and patient and public engagement can meaningfully complement such an opt-out model and help ensure the ethical use of samples and data.

Paper: Dynamic consent: a royal road to research consent?

Authors: Andreas Bruns and Eva C Winkler

Affiliations: Section of Translational Medical Ethics, National Center for Tumor Diseases, Department of Medical Oncology, University Hospital Heidelberg, Heidelberg, Baden-Württemberg, Germany

Competing interests: None declared