28 August 2019
Share this article

Navigating the moral maze of open-ended data collection

On June 6, 2019, Martin Boeckhout, Senior Consultant at the MLC Foundation, successfully defended his PhD thesis on ‘The Ethics and Governance of Health Research Infrastructure’. It not only gained him his PhD, it also represents an important contribution to current and future debates on the Ethical, Legal and Social Implications (ELSI) of biobanking, big data analytics, data portability, governance structures and the inter-relationship between research and healthcare in the Netherlands.

“Biomedical research infrastructures face three important ethical challenges over the coming years. The first is related to the open-ended use and storage of data. Currently, we don’t really have criteria for assessing whether that is done appropriately. The second challenge is the trend towards closer integration of research data collection into routine clinical care. Ethically speaking, clinical trials traditionally required separation between trial participation and the administration of healthcare, but for practical reasons and in terms of basic principles of operation that doesn’t work for large-scale research infrastructures,” explains Boeckhout. “The third challenge relates to patient and participant engagement, because the previous focus of ethics on what clinical trial participants are subjected to doesn’t fit the scenario of large-scale data collections. Today, we’re collecting data from people who are not physically being subjected to interventions, yet as active participants they still expect to have a say in what research based on their data should be about.”

Data Stewardship
Today’s large-scale biomedical databases and biobanks are founded on two important premises. Firstly, that the most effective way of increasing the statistical power of data analysis is to increase the sample size, and secondly that solving complex challenges in medicine requires a highly collaborative multidisciplinary approach. To help research groups find the data they need, much has been publicized about the need to make data FAIR – Findable, Accessible, Interoperable and Reusable. However, as Boeckhout points out in his thesis, from an ethical standpoint, implementing the FAIR data stewardship guidelines won’t solve all the problems of data sharing.

“Two assumptions are often made about FAIR data stewardship. One is that more data sharing is always better. The other is that if you prepare data for sharing according to the FAIR principles, then sharing will actually occur. However, these assumptions aren’t necessarily correct,” comments Boeckhout. “Whether more sharing is actually better depends on how data infrastructures and the uses to which they are put are governed. The FAIR principles alone don’t address this. Simply offering greater opportunities to share is not going to resolve the general issue that the reward system in science remains too focused on research excellence and insufficiently focused on quality, patient needs and the fair distribution of rewards.”

The division between research and healthcare
Conventional clinical trials have always been characterized by a recognizable division between research and healthcare, with informed consent mechanisms tailored to this purpose. However, as Boeckhout points out, that doesn’t translate very well to large-scale biobanks and data registries.

“When setting up large-scale biomedical databases and biobanks, there are issues over how to collect data and tissue in such a way that patients are not overburdened by the process. That’s often cited as one of the reasons for integrating research data collection into standard healthcare protocols and workflows. But the problem is that if you integrate it very well, it also tends to become hidden from view. And the question then becomes – are patients well enough informed about the potential research that will be made from it? It is important that patients always have the choice not to participate in research,” says Boeckhout. “We need to rethink the role that ethics review can play in such circumstances – do we stick with independent ethics review in advance of projects or do we add some form of ongoing ethics assessment in the governance structure of major programs.”

Another issue discussed in his thesis is that of incidental or secondary findings – anticipated or unanticipated findings that are outside the original purpose of the research, or additional results actively sought by researchers.

“As data analysis becomes more ubiquitous, for example, when whole genome sequencing becomes routine, the potential for incidental and secondary findings becomes much greater. This raises serious ethical issues. On the one hand, for instance, patients have a right not to know about such findings, but on the other hand clinicians have a duty to act in the best interests of patients and research participants. But they also have to keep in mind that in a research setting their data quality might not be as high as in a clinical setting,” he says. “I think one of the main issues here is how to deal with uncertainty and risk, not just from the perspective of research and assessment of health consequences, but also how to deal with risk and uncertainty when communicating findings to patients.”

Patient participation and engagement
While there are ethical issues around how you communicate research findings to patients, there are also ethical issues around recruiting patients and participants in the first place.

“More often than not patients would say they are happy to make a contribution to research by contributing their leftover tissue, given that their tissue might otherwise be wasted,” says Boeckhout. “But if you ask ordinary citizens if you can use their personal data for research purposes, they have a much more differentiated perspective. When approached as citizens, the considerations become more political, and debates on privacy and data sharing become more charged. My opinion is that people in general would like to contribute their data to research, but as citizens they are also thinking about what they consider good use of their data and who they trust with it. Then all kinds of other considerations come into play as well, like for example, commercial use. Unless these issues are confronted and properly addressed, you could easily find that people lose trust in health research infrastructures.”

However, Boeckhout believes that a number of forums for debate and suitable mechanisms for implementation are already in place. His thesis, for example, includes guidelines developed as part of a BBMRI-funded project he conducted about how biobanks might be able to get patients and participants onboard more actively.

“That project eventually led to me and Gerhard Zielhuis being given the opportunity to set up BBMRI-NL’s Patient and Public Advisory Council, which I am happy to say is expected to become adopted by Health-RI,” he says. “In general, I think the Netherlands has a vibrant and diverse research infrastructure community which includes many patient advocacy and participant voices. There are already many experienced people and ELSI researchers out there that anyone setting up a new biobank or data registry can learn from.”