Broad consent for biobanks is best – provided it is also deep

BACKGROUND: As biobank research has become increasingly widespread within biomedical research, study-specific consent to each study, a model derived from research involving traditional interventions on human subjects, has for the sake of feasibility gradually given way to alternative consent models which do not require consent for every new study. Besides broad consent these models include tiered, dynamic, and meta-consent. However, critics have pointed out that it is normally not known at the time of enrolment in what ways samples deposited in a biobank may be used in future research and that, for a consent to be informed, exactly this kind of knowledge is required. Therefore, there is an ongoing debate about the ethical acceptability of going for less than study-specific consent. MAIN TEXT: In light of this debate we address the question of how to best protect participants against relevant risks and violations of autonomy. We apply the central aims of the informed consent process to the unique circumstances of biobank research where samples and data in many cases are stored for long periods of time and reused in subsequent studies. Thereby we are able to formulate a set of criteria focusing both on the risk of informational harm and the potential violation of participants' values. We compare existing models of consent based on their ability to satisfy the criteria, and we find that the broad consent model offers the best level of protection for participants, although, it suffers from a few important deficiencies with regards to protection against participant value violations and long-term protection of autonomy, if it is applied without qualifications. For this reason, we propose modifications to the current broad consent model, in order to ensure that it provides protection of autonomy and participant values through strong ethical review and continuous communication. CONCLUSION: We conclude that a modified form of broad consent is ethically superior in biobank research, not only because it is most feasible but primarily because it offers the best available protection against the hazards facing research subjects in this form of research.