“The worst possible thing we could do is give data away to a company that doesn’t pay tax in the UK, finds a cure for dementia, then decides to sell it to us at a price we can’t afford,” says Annemarie Naylor, ex-director of policy and strategy at Future Care Capital (FCC), a charitable health and social care think tank.
The NHS holds 55 million primary care records and 23 million secondary and tertiary care records along with many other scientific and medical datasets. John Bell, the Oxford University professor who leads the government’s Life Sciences Industrial Strategy, has said that the innovations possible from using this data are “mammoth” and will “change medicine forever”.
The task for policymakers is to avoid Naylor’s worst-case scenario. There are two broad challenges. One, convincing a reluctant public that their health data can be used safely and securely to, in Naylor’s words, “secure the innovations we all want for our grandchildren”. And two, developing a robust value framework for this data so that, in deals with the third parties capable of developing such innovations, the NHS can receive fair and equitable returns for the data it supplies.
The widespread opposition that forced the government to delay its centralisation of GP data earlier this year demonstrates that little progress has been made in convincing the public that its health data can be used safely and securely. Recent deals with third parties also show the struggle to secure fair returns for the NHS – a 2016 collaboration between Moorfields Eye Hospital in London and Google’s AI division DeepMind generated an algorithm to improve the diagnosis of age-related macular degeneration, a condition that can lead to blindness. But while Moorfields gained the algorithm, DeepMind retains the intellectual property and the right to all the returns if the algorithm is sold worldwide.
Tackling the challenge of public distrust is vital because, according to Naylor, “there is no data asset without trust” as people must consent for data to be curated and explored. While safeguards governing data access and use have certainly been strengthened, especially with GDPR, polling has shown that the public expects these safeguards to be in place. Public distrust is sustained by how they believe third parties will use their data – a huge 78 per cent think companies use personal data for their own benefit, not for the public’s.
Addressing the issue of public benefit could, therefore, tackle both public distrust and help secure returns for the NHS. This is precisely the aim of the Sovereign Health Fund policy developed during Naylor’s time at the FCC.
The idea is big and would be “truly pioneering”, she says. A Sovereign Health Fund would be a “portfolio of data-driven investments that would recoup some or all of the value that we felt was tied to that data”. NHS data would be curated and placed in trusted research environments (TREs) around the country; third parties could apply to use the data to generate innovations, ensuring that a certain amount of any innovation’s returns flow into the state-owned fund and are ring-fenced either for investment into the health service or further research.
Naylor is clear the system would directly address concerns about public benefit by “generating a return from a very key national asset [health data] rather than [it be subject to] asset stripping”, and adds that “people would know it would be reinvested in social care and the NHS”. No part of the NHS would move into private hands. It doesn’t have the capacity to generate these innovations alone, so the policy would facilitate third-party collaborations and secure fair returns for the public health service.
A key step for delivering the policy would be developing a value framework setting out how NHS data would be valued in deals with third parties. Despite making positive noises on the creation of TREs and the possibility of a life sciences fund, the government hasn’t yet set a timeline for publishing its value framework.
However, valuing data is not easy – there are still no agreed international standards in accounting terms. Part of the pioneering character of the policy would be changing this. It would be necessary to have a period in which government builds the market for this data by working with third parties to assess its commercial value and understand what sort of innovations would be possible.
If a value framework was established and such a market made, NHSX, the health service’s digital and data arm, would play a key role. In 2020, the Centre for Improving Data Collaboration was created within NHSX “to provide specialist advice for health sector organisations entering into data partnerships”. The centre would need to use the value framework to advise on deals that may vary hugely – a partnership with a university researcher may only need to cover the cost of supplying the data, while a partnership with a large, profit-seeking corporation would require a deal that secured fair returns for the NHS data provider, unlike the Moorfields-DeepMind deal.
The use of NHS data represents one of the biggest political challenges but also one of the biggest national opportunities of the coming decades.
As a policy answer, a Sovereign Health Fund has the potential to make data partnerships between the NHS and third parties both safe and hugely beneficial for future innovation and funding of the health service. What it will take from government is ambition but, above all, transparency. This policy can only gain support if the public is confident in the safeguards around their data and confident that fair value will be recouped and reinvested in the country’s health services.