select search filters
briefings
roundups & rapid reactions
before the headlines
Fiona fox's blog
Experts react to news of Palantir operating the NHS’s new federated data platform.
James Davenport, Hebron & Medlock Professor of Information Technology, University of Bath said:
“This is much more limited development than the rumours in the computing press. It would seem to be (but the NHS needs to vigilant here) about vacancy data and planning, rather than the detailed health of individuals. The main threat here is what the military call “mission creep”, as Palantir etc, say “we could do a better job if we had access to a bit more data”.
Professor Peter Bannister, Managing Director, Romilly Life Sciences and Healthcare Policy Chair, Institution of Engineering and Technology, said:
“With demand for NHS services continuing to outstrip demand, widely deployed technology solutions that are enabled by data remain a must-have if we are to close this gap.
However, to achieve the greatest impact it is essential that use of the Federated Data Platform is rigorously and transparently monitored to build confidence amongst patients and healthcare professionals so that its use becomes routine.
We also need to understand that this is one step in the journey towards truly preventative healthcare which will ultimately require the integration of GP data, as well as information from sources such as wearable devices, to deliver sustainable and cost-effective healthcare for all.”
Professor Elena Simperl, Professor of Computer Science at King’s College London, said:
“The NHS is probably unique in the world with respect to the data they hold, which some have valued to close to £10 billion. That data is extremely valuable, to improve how the NHS operates, to drive advances in scientific research, and to boost AI products and services. There is a huge responsibility in stewarding and governing this data asset, not only for these reasons, but only because every failure to maintain and restore public trust could inevitably lead to more people opting out of secondary uses of the data, hence diminishing the benefits the data could yield. We have experienced several attempts over the past years to achieve similar aims – no matter how well intended, they fail because of a common set of concerns around data privacy, stewardship and governance and it is unclear to me whether lessons have been learned from these experiences. With every failed attempt, more people have opted out, and public trust has been eroded.
“Attention to data is key to every large organisation and there is a lot of experience in the data community around the principles and practices that lead to healthy data ecosystems for the public good. These include the use of open approaches (open source, open science, open standards), which do not compromise on data privacy and have built in transparency, accountability and community-led quality assurance and maintenance features.
“Besides the principles and technical capabilities of the federated data platform, it’s clear that a lot remains to be done to improve communications around the sharing and use of sensitive data and advance our thinking of appropriate structures for stewarding and governing public interest data to regain and maintain public trust.”
Professor Jon Crowcroft, Marconi Professor of Communications Systems, University of Cambridge, said:
“My question would be why give this to anyone given the UK Health Data Research has already federated processing across all the main centres for health research, and doing it for patient services operationally is much easier – the DARE program addresses ethical, legal and policy challenges and fixed all of them and now has technology out there – it’s a UK funded program and only uses UK funded people – why give money to a sketchy overseas outfit with known behaviour that is not aligned with NHS ethos? See https://www.hdruk.org/access-to-health-data/uk-trusted-and-connected-data-and-analytics-research-environments/ for the existing setup.”
Professor Andrew Morris, Director of Health Data Research UK (HDR UK), said:
“The highest performing health systems internationally have whole system intelligence – the ability to follow journeys of care of patients to ensure care is person-centred, safe, effective, efficient, equitable and timely. This is underpinned by data.
“It is vital that data is used well in the NHS as it supports the best treatment and care by ensuring complete, secure and up-to-date information is available to healthcare professionals at the right time. It’s what we expect as patients.
“The NHS intends the Federated Data Platform to enable better use of our medical data in the care we receive – to join up data that may be held in different parts of the NHS to create these journeys of care, and to help see where the service has capacity for offering treatments we may need.
“But there is a risk. We are worried that concerns about the platform may lead people to use opt-outs that will have no impact on the Federated Data Platform but will harm important medical research.
“Public confidence and demonstrating trustworthiness is crucial. If people have questions about the FDP that they feel aren’t being answered, or there are some that want to protest the move, the National Opt-out Scheme is likely to be used as a route to express this. But NHS England states that opt outs would not apply to the FDP. Opt outs will certainly and directly impact the completeness of data used in research, which would reduce our ability in the UK to carry out important research to improve the treatments and care that we all, our families and our loved ones receive.”
Notes:
The NHS has provided an FAQ to answer some of the questions the public might have about the procurement process, who will be able to access their data, etc https://www.england.nhs.uk/digitaltechnology/digitising-connecting-and-transforming-health-and-care/fdp-faqs/#what-is-the-federated-data-platform-fdp
Nicola Perrin MBE, CEO of the AMRC (Association of Medical Research Charities), said:
“It is crucial that the public, patients and healthcare professionals have confidence in the way NHS data is used. The FDP is incredibly complex so having clear and accessible information about what it is and is not, is really important, for example it is about care and care services not research. We are pleased that NHSE has created an engagement portal to coincide with their announcement.”
“The overarching goal to make better use of data to help the NHS improve how it arranges and delivers care for patient benefit is a worthy one. The creation of the FDP must become part of an ongoing dialogue about the use of health data*.”
*https://www.amrc.org.uk/blog/where-were-at-with-the-federated-data-platform-fdp
Professor Carissa Véliz, Associate Professor at the Institute for Ethics in AI, University of Oxford, said:
“The devil will be in the detail, but if the NHS gives Palantir —an espionage American tech firm— any access to sensitive data, which seems likely, it will undermine public trust yet a little bit further. Public health systems do not work when there is no public trust left, and the NHS is running out of it”
Professor Meg Davis, Professor of Digital Health and Rights at the University of Warwick, said:
“Palantir has a very concerning history in the US of expansive surveillance and targeting, in ways that are intrusive and could undermine the rights of vulnerable groups. Any agreement with Palantir should be made public in the interests of transparency, with tough privacy protections.”
Professor Oli Buckley, Professor of Cyber Security, University of East Anglia, said:
“In a move poised to revolutionise patient outcomes, the centralisation of medical data emerges as a double-edged sword, offering unprecedented opportunities in healthcare while simultaneously ringing alarms about privacy and autonomy. Palantir, with its proven prowess in navigating vast and varied datasets, stands at the forefront of this initiative. Yet, the company’s entanglements with intelligence agencies cast a long shadow over its role in handling sensitive health data. This partnership raises the spectre of Big Brother in the realm of personal health, leaving many to grapple with the implications of surrendering their medical histories to a watchful, all-knowing guardian. As we stand at this crossroads, the critical question remains: At what point does the promise of better healthcare begin to erode the foundations of our privacy?”
Professor Stephen Evans, Emeritus Professor, London School of Hygiene and Tropical Medicine, said:
“ Currently different hospitals use a variety of different computer systems to handle patient data used for clinical care. Quite often there are several different systems used within a single hospital by different departments. This means that for an individual patient who uses hospital facilities, including things like laboratory tests requested by GPs, the data related to that individual patient may be on several different computer systems. For the doctors and other healthcare staff this may mean that important information about that patient may be difficult or even impossible to access if it is on a different computer system.
A ”Federated Data Platform” is another computer system that, as it were, “sits on top of” all these different systems and is able to access all the relevant information for a patient so that a health professional can see it all. The current proposal relates only to hospital computer systems and does not include data stored on GP computer systems.
In principle, making fuller information available related to the care of the individual patient must be beneficial. This proposal is not replacing all the different systems with a single system, but allowing for data from different systems to be accessible using a single system. It seems likely that this will be gradually rolled out across the NHS and will not happen immediately.
There are obvious concerns that privacy may be less protected than it is now, but those concerns may not be well based. However, handing the control of such a system to a company interested in short-term profit and whose founder seems to believe in the abolition of the NHS is a risky strategy. By its nature it may prove more difficult to implement the single view of data than is envisaged, even though the aims are laudable. It has the potential to absorb large amounts of NHS money, especially in the costs of “tweaks” to the system after it has started. Overall, it may or may not lead to improvements in the efficiency of patient care.
It is likely to be used to produce summary data around patient care and could be used to carry out research utilising data from many patients. If this is done it is vital that a “Trusted Research Environment” (TRE) is implemented which has a very different objective to the “Federated Data Platform” (FDP). The FDP is dependent on identifying the individual patient, whereas the TRE, while using individual data, ensures that individuals cannot be identified, and has the building and retaining of trust as a major objective. Both FDPs and TREs have their uses but it may lead to “opt-outs” from use of data for research that will not affect the FDP since opting-out for direct care is not possible, but it would severely damage other health research carried out in a TRE.
Good quality research requires careful analysis using a TRE. The potential for poor quality research derived by naïve use of summary data is considerable with the FDP. It is vital that trust is maintained in the processing of patient data so that research, such as that done during the Covid pandemic which identified vulnerable groups, is able to continue.”
Declared interests:
James Davenport: No COIs
Peter Bannister: “Honorary Chair, University of Birmingham. Non-exec Director, Life Sciences Hub Wales”
Nicola Perrin: “Part of NHS England’s check and challenge group for FDP.”
Andrew Morris: “No conflicts of interest, except the mission of HDR UK: to unite the UK’s health data to enable discoveries that improve people’s lives.”
Oli Buckley: No COIs
Stephen Evans: “SE was involved in the OpenSafely Collaboration which used a TRE based on GP data to produce many research papers during the Covid Pandemic.”
For all other experts, no reply to our request for DOIs was received.