Should biobanks release diagnostic data to patients?

The UK Biobank aims to improve the prevention, diagnosis and treatment of chronic diseases for future generations by tracking the health and wellbeing of 500,000 volunteers to understand why some people develop certain diseases while others do not.

Information about these volunteers is provided to approved researchers, both in the UK and overseas, from academia and industry. The participants, who were all aged between 40 to 69 in the years 2006 to 2010, have provided blood, urine and saliva samples alongside detailed information about themselves and agreed to have their health followed.

A recent University of Exeter study of 201,465 blood samples from the UK Biobank identified 2,022 as-yet undiagnosed cases of type 2 diabetes. The patients found to have undiagnosed diabetes waited 2.3 years on average following UK Biobank recruitment for a clinical diagnosis. Five years on, 23% still hadn’t received a diagnosis. These participants were typically older, more likely to be obese and more commonly male than those who did not have diabetes.

Volunteers taking part in the UK Biobank project consent to not receiving the results of any tests carried out on the samples they’ve provided to the project – this particular study, which sought to understand the prevalence of undiagnosed type 2 diabetes in the adult population, outright hinges upon them not finding out. The identification of these patients gives researchers a unique opportunity to study the impact of delayed diagnosis on the risk of developing complications in the future.

Information from UK Biobank findings isn’t always kept under wraps. Participants in the UK Biobank’s MRI Imaging study are put through a number of routine scanning procedures and actively informed of any potentially serious findings. Benign findings, however, are not disclosed.

While patients can probably get away with finding out about things like gallstones and arthritis later rather than sooner, not telling to somebody that they’ve lived for several years with a condition, such as type 2 diabetes – which is known to shorten patients’ lifespans by up to a decade – is an ethically fraught decision. It’s a grim spectre that hangs over biobank projects worldwide, where medial research butts horns against the better interests of individuals’ health.

While outright disclosing all recorded heath data to participants as soon as they enrolled in the biobank would be counterintuitive to its very purpose, one might argue that there should be a set of conditions imposed where participants can be advised that they’ve tested positive for certain conditions.

For instance, should they be found to have a condition like type 2 diabetes, the biobank could inform their primary care physician after a suitable time period has passed – one long enough to allow researchers to assess the benefits of earlier diagnosis, but not so long that it could be substantially detrimental to a person’s long-term health.

This would allow the UK Biobank to continue its vital work, without meaning that people are left in the dark about the state of their own health data.

For all the vast potential that biobanking and big data brings to medical research and the prospective of predictive healthcare, it also throws up a number of ethical quandaries, from data anonymisation and privacy issues to the complexities of third-party data access and consent.

One of the thornier dilemmas for biobank ethics committees to consider is whether, and when, to share incidental findings that might suggest undiagnosed health issues or risks with individual participants.

It’s an ethical conundrum – does the duty of care between a biobanking project and a participant include an obligation to disclose potentially life-altering medical information about which the individual didn’t already know?

In the case of the UK Biobank, participants consent that they will not receive feedback of individual results, and for good reason – the long-term data around outcomes and patients’ interactions with their health systems would be spoiled if biobanks (or the third-party researchers running studies on them) intervened to reveal new medical information to participants.

In the case of the aforementioned diabetes study, researchers have the opportunity to study the impact that delayed diagnosis might have on later-life complications and draw conclusions that could drastically improve public screening and treatment of the disease.

There is a wide-ranging debate around whether certain pieces of medical data retrieved from biobanks should be shared and, if so, in what way. There are perfectly valid arguments for doing so, and consent templates have been proposed to define the criteria to judge when a disclosure may be justified. But in any case, extreme caution should be used, because the implications of releasing such data are complex.

The UK Biobank instituted its own protocols for disclosing potentially serious incidental findings during a major, ongoing imaging analysis of 100,000 samples. The system involved radiographers flagging potentially serious incidental findings from medical images, which were then referred for confirmation by a specialist radiologist before being passed on to participants and their GPs.

A 2018 follow-up analysis of this process found the system to be broadly effective at minimising harm, but still highlighted the downsides of sharing such data, concluding that “the great majority of these findings turn out not to be serious, resulting in unnecessary anxiety for the participant and unnecessary clinical assessment, which may include invasive procedures provided by publicly-funded health services”.

Unnecessary anxiety for participants and wasted costs for health systems may be balanced by the cases in which serious health conditions were diagnosed earlier than they would have otherwise. Nevertheless, it highlights the risks involved in sharing incidental findings, and the burden involved in managing the knock-on effects for projects that are meant to support foundational research, not provided individual diagnoses.

While the debate over consent and data-sharing from biobanks continues, it’s important to remember that biobank data is of greatest value when used for its original purpose – to inform health strategy, including the optimal targeting of public screening and clinical diagnostic checks, rather than acting as a source of diagnosis in itself.