The ethics of consent for the public release of potentially identifiable DNA data

A major ethical and policy challenge facing genomics research stems from the existing mandate for rapid public release of all sequenced DNA data. It is now clear that an individual can be uniquely identified with access to a small number of SNPs from that person. Genome-wide association studies routinely use more than 100,000 SNPs to genotype individuals, creating privacy risks that are only going to increase as technology advances and electronic databases proliferate. Informed consent is not currently required for DNA data release because de-identified data are incorrectly assumed to be unidentifiable. We have argued for policy reform that mandates informed consent for data sharing. The goal of this proposal is to spearhead this effort by developing a consent process that will encourage data sharing while building public trust and fostering participation in genetic research. This project builds on a preliminary study of participants' attitudes toward DNA data release and has three specific aims: (1) conduct a randomized trial of three alternative types of consent for DNA data sharing (traditional, binary, and tiered) to compare their impact on enrollment and consent to data sharing, (2) evaluate subjects' judgments about data sharing and assessments of traditional, binary and tiered consent, focusing on calculations of risks and benefits, informational needs, and desired levels of control over decision making, and (3) develop clinically relevant policy recommendations for DNA data release. Participants in ongoing genomic studies of cancer and epilepsy at BCM will be invited to participate. Responses will be compared by type of consent (traditional, binary, tiered), subject population (patient, parent of affected minor, control), and disease type (epilepsy, cancer, healthy volunteer). It is our hypothesis that tiered consent will be best able to satisfy the range of participants' judgments, will yield the highest consent to data sharing, and will not negatively impact enrollment into genetic research. This project will culminate in a practical model consent process and form for data sharing that can be adapted by investigators and used as a guide by IRBs, and the results will help shape a broader conceptual model for ethical variation in types of consent for human subjects research. Moreover, it will increase our understanding of participants' judgments and attitudes toward data sharing, which will improve the overall conduct of biomedical research, build public trust, and foster research participation.