This morning’s discussion – jointly hosted by the All-Party Parliamentary Groups on Data Analytics and Health – suggested that if we want uses of health data to be trusted, we need to trust citizens and patients to think more deeply about benefits and risks than media headlines might suggest. The session was inspired by a recent Understanding Patient Data study and report which used citizen juries and a survey to discover public attitudes to data partnerships between the NHS and researchers, charities and businesses.
The results indicate that both citizens and patients are keen to engage, and can quickly develop sophisticated understandings of benefits and risks in data partnerships. If we give them the chance…
The framing of the study must have helped, by asking what would a fair partnership look like? This encouraged contributors to range widely in their interpretations: fair to individual patients whose data may be used; but also fair between the health service and its current and future partners (benefits should not be locked away to commercial advantage), fair to the population (benefits should not be localised, either geographically or by community), fair to past patients (who almost always want their experiences to benefit others), and fair to future patients (who should be able to benefit from the experiences of those who have been treated before).
But opportunities to benefit from partnerships are not limited to patient data: providing health services also generates large amounts of administrative and operational data, and a growing amount from sensors. For example data has already been used to route supplies of blood products, which have a very short shelf-life, to places where they are most likely to be needed.
Since the study was completed just before the onset of the COVID-19 pandemic, panellists considered what had changed. Public awareness of the role of data has definitely been raised, with both good and bad experiences being explored by the media in more detail than might have previously been the case. COVID-focussed research has revealed surprising, and important, gaps in data: 26% of Electronic Health Records (EHRs) omit ethnicity, making them hard to use to determine whether this is a factor in the severity of the disease; long COVID is poorly detected by any quantitative statistics, and EHRs are not a good vehicle for capturing the qualitative data where this form shows up. Responding rapidly to the onset of the pandemic meant that a number of partnerships were set up quickly: as has been suggested for education we should be reviewing the policies, infrastructures and norms that those emergency measures encoded to ensure they are appropriate to maintain public trust in the longer term.
All of which seems to make the study’s recommendations even more relevant. Health is such a sensitive issue that trust will always be fragile and need to be continually re-earned. This requires good governance, accountability and meaningful transparency (see below). There is an expectation that citizens will be involved in decision-making, both about uses of personal data and anonymised, and an appetite to understand how data can be used for public benefit. Since it should also provide a wider perspective to reveal unexpected consequences, early citizen engagement should be a win-win. Even if full public consultation is not possible in a rapidly-changing situation, having citizens/patients as regular members of boards, panels and juries provides many benefits.
Communications are likely to be critical to trust in any partnership, and need to be developed alongside technology and health, perhaps even before. This is a specialist task, particularly given the need to address many audiences, who may want a simple reassurance or a deep dive. We need to be ready to invest in expert support, especially when circumstances demand communication needs be done quickly. We must be open about benefits and risks: citizens have now seen sufficient unsuccessful data projects to presume the worst if the latter are not mentioned, or if information is not available. But we should also celebrate the successful partnerships that have been going on for many years, perhaps not wishing to draw attention to themselves for fear of adverse publicity.
Communication and support is also likely to be needed within the health and research communities: if expectations, safeguards and red lines are not explained and understood then cautious interpretation is likely to be much more of a limit than law, policy or governance actually require. We need to get better at explaining what rules such as the GDPR already allow, why and under what safeguards, rather than letting them be cited as barriers. Here the virus response may provide a model, with simple Information Governance guidance being issued quickly, and cross-skilled teams assembled to work together on technical, legal, governance and communications requirements.
It seems to me that much of this is applicable far beyond the health sector. Especially the closing line: “regulation and governance are not barriers to responsible innovation, they allow it to go faster: like brakes on cars”.