Last week the National Institutes of Health released its genomic data sharing policy. From next year, study participants will need to give researchers permission to re-use and share their data. But, what about many older studies that were obtained years before re-use of samples was envisaged? Michele Cote and colleagues from Wayne State University and the Karmanos Cancer Institute, USA, wondered what would happen if people from a 15-year old lung cancer study were re-contacted and asked to share their de-identified data.
Fifteen years ago when I embarked upon a career in medical research as a project coordinator, the world of high-throughput genomics was in its infancy. The Human Genome Project was nearing completion, but most genome research taking place at the time was focused on a handful of genes (“candidate genes”), with one or two single nucleotide polymorphisms (SNPs) of interest in each gene.
As the coordinator of a population-based lung cancer study, my main focus was to recruit hundreds of patients and suitable controls, to obtain informed consent from the study subjects, and to coordinate interviews and DNA collection. The idea that in the future the DNA from these participants, many of whom were terminally ill, would be sequenced to discover nearly all germline variation was far from my mind.
Fast forward to 2008. The National Institutes of Health (NIH) started to require NIH-supported studies and submit de-identified data generated from high-throughput genomic analyses to the database of Genotypes and Phenotypes (dbGaP). Most would agree that this is for the greater good, because maximizing the utility of these data for biomedical research is beneficial and should be promoted. Maintaining the confidentiality of these data was recognized as crucially important; however, it was also recognized that most informed consent documents from studies prior to this requirement did not explicitly allow for release of these data.
The high cost and time it takes to collect new study subjects, especially when adequate DNA remains in storage from participants already consented to genomic analyses, made us wonder: If contacted again to request permission to release this information, how would people respond? Our research team agreed that releasing data to dbGaP was beyond the scope of the original consent form and that re-consenting patients was the ethical thing to do.
At our institution, we identified nearly 2,500 people who had participated in one of three lung cancer studies that had taken place in metropolitan Detroit from 1999-2009.Many participants (25.3%) were already deceased, leaving 1,833 people to re-contact to gain their consent to release their data to dbGaP.
First, we attempted to contact people through the U.S. postal service. We sent an introductory letter and consent forms by mail, along with a postage paid return envelope. Next, if the letter failed to generate a response after several weeks, we called by telephone. On average, 3.5 phone calls were made to each participant. We were never able to reach 37% of the individuals we attempted to re-contact. Of those we did have contact with, the majority agreed to the release of their de-identified data, yet 11.3% of individuals did not want to release their information to dbGaP. Race, gender, and age at initial study enrollment were not associated with whether or not an individual agreed to release information.
No consensus has been reached by the scientific community regarding how to best inform study participants about the potential future uses or sharing of their data and biological specimens. Most prospective studies included many years of participant recruitment and often at the time a patient agrees to donate a specimen, we can’t imagine what technology will exist a few years down the road.
Given the collaborative nature of scientific research, the decreasing costs of genomic analyses, and the challenges associated with participant recruitment; the sharing of retrospectively collected data and specimens is needed to advance science. This will present ethical challenges to investigators in ways that today we have yet to imagine. We need to continue to investigate how to balance the use of these data and biospecimens in a way that protects the interests and confidentiality of our participants, while still allowing the rapid advancement of science.
Dr Michele Cote is an Associate Professor at Wayne State University and a Scientific Member of the Karmanos Cancer Institute.