What did the Guardian find?
On 2 November 2026, The Guardian published an investigation by Robert Booth into the NHS Federated Data Platform (FDP). The findings emerged from more than 100 Freedom of Information (FOI) requests. Medact, a medical charity that opposes the contract, gathered the responses and shared them with the paper.
The central finding was narrow and striking. Twenty-nine hospital trusts are on NHS England’s list of organisations “realising benefits” from the platform. Those trusts said it was not live, or that they were not using its applications. The trusts are not anonymous. Epsom and St Helier said it was “not using FDP at the moment”. Kent and Medway said the system was “not currently live”. Mersey Care said it had “not yet gone live”.
Nottingham University Hospitals said “no implementation timeline has been agreed”. University Hospitals of North Midlands was more blunt: “FDP is not live. We do not intend to roll it out in the future.” Those statements sit awkwardly beside a public list that presents the same organisations as evidence of the platform’s value.
The paper says more than two dozen trusts were “not using all or most of the AI-powered technology”. Its detailed count is 29. Several of those trusts said the platform was not live at all. The contrast is between a national list that marks trusts as benefiting and local statements that the system is absent or idle.
NHS England disputed the implication. It told The Guardian that nearly 140 trusts “have now reported benefits” from the platform’s tools. It also said its list includes trusts that use the FDP for back-office analysis and data submission, not only for patient care. Some trusts that say they do not use the platform may still use a small part of it to calculate treatment costs.
Palantir, the US company that leads the supplier consortium, pointed to its own measures of success. It cited additional operations and a 14.3% reduction in discharge delays for long-stay patients. It also cited a 5.6% improvement in cancer diagnosis within 28 days. NHS England’s published figures use slightly different numbers, which this article sets out below.
What is the NHS Federated Data Platform?
The FDP is a software layer that sits above the NHS’s existing databases. It is meant to let hospitals, integrated care boards and national bodies connect operational data that is stored in different systems. Each participating organisation has its own instance, and NHS England runs a national instance. Palantir describes the platform as a way to organise information held in separate hospital databases.
The NHS has no single central database of patient records. Hospitals store operational data in different electronic record systems, waiting-list tools and finance systems. The FDP is meant to create a unified view without copying everything into one central store. NHS England says each organisation remains the data controller for its own instance.
The contract was awarded in November 2023 to a consortium led by Palantir Technologies UK. Accenture, PwC, NECS and Carnall Farrar are also involved. The contract began in March 2024 and is worth £330m. It has an initial three-year term, with options to extend it to a maximum of seven years.
The platform is intended to support waiting-list management, theatre scheduling, patient discharges and cancer pathways. It also handles back-office reporting and data submission. NHS England says 142 of 202 NHS trusts were live on the platform at the end of August 2026. A total of 170 trusts had signed a memorandum of understanding to join. Thirty-five of 36 integrated care boards were live.
Use is not mandatory. NHS England clarified that point to local organisations in 2026. The original target was for 85% of secondary care trusts to adopt the FDP by March 2026. NHS England’s planning framework says all providers should be onboarded by 2028/29 and use its core products.
Supporters say the platform can help staff find information in one place. Critics say it adds another layer to systems that already struggle to interoperate. Both descriptions can be true at once, which is why the definition of “use” matters.
Why does 'realising benefits' matter?
The public debate often treats “live”, “using” and “realising benefits” as if they mean the same thing. They do not. A trust can be listed as live because it has a local instance of the platform, even if clinicians rarely open it. A trust can report benefits because a small number of staff use a data-submission tool, even if the clinical apps are dormant.
Torbay and South Devon NHS Foundation Trust illustrated the gap. It told The Guardian it was “not currently operationally using the FDP”. NHS England still records it as live. The infrastructure is in place “to support data submission, rather than active operational use of FDP apps”. That is a defensible definition for a data programme. It is not the same as a hospital using AI to treat patients.
This distinction matters because the benefits list is the strongest claim in the programme’s public case. NHS England’s page says 137 trusts were “reporting benefits” at the end of August 2026. That figure is used to justify the rollout and the contract. If 29 of the trusts on that list say the system is not live or unused, the claim needs a clearer definition.
The patient-care product counts are much smaller than the headline. NHS England says 40 trusts were realising benefits from the inpatient theatres module. Another 37 were realising benefits from the waiting-list module, and 31 from the referral-to-treatment tool. Thirty trusts were realising benefits from OPTICA for discharges. Six trusts used Cancer 360, and five reported benefits from it.
Those numbers are not hidden. They are published product by product, below the headline totals. But the headline is what politicians, suppliers and reporters tend to quote. The product-level data gives a more accurate picture of how deeply the platform has changed clinical work.
The gap also affects how the programme is judged. A trust that has connected its infrastructure may count as a success for adoption. A trust that has not changed clinical work may count as a failure for patient care. Both statements can be true at once. The argument is about which one the public is being shown.
How strong is the evidence behind the FDP's benefits?
NHS England publishes two kinds of benefit. The first is a count of actions taken by local users, such as records reviewed with a waiting-list tool. The second is a before-and-after calculation of operational performance. The second kind produces headline numbers such as additional operations or shorter discharge delays.
The published figures are substantial. NHS England counts 117,539 additional patients who underwent procedures in theatres. The cumulative figure runs to the end of June 2026 and compares with the period before FDP use. It reports a 14.34% decrease in average delay days for long-stay patients after OPTICA was introduced. It also reports a 5.68% average improvement against the 28-day Faster Diagnosis Standard for cancer.
Those figures come with a caveat. On 6 June 2026, NHS England added a sentence to its methods page. It said: “We cannot therefore draw conclusions about cause and effect as other variables have not been controlled for.” The Office for Statistics Regulation (OSR) reviewed the change and the surrounding communications. It welcomed the caveat and said NHS England had agreed to label before-and-after comparisons more clearly.
The OSR listed specific commitments. NHS England agreed to add a caveat to the main benefits page. It also promised a schedule for trust-level data. Case studies would be labelled as written by organisations using the FDP, not verified by NHS England. The regulator said it would monitor how ministers and official publications use FDP figures.
NHS England has also commissioned Imperial College to carry out an independent academic evaluation. That evaluation is expected after the government decides whether to extend the contract. The National Audit Office is separately examining how the Department of Health and Social Care and the NHS are integrating patient data. Its report is expected in winter 2026/27.
The timing is awkward. The independent evaluation will arrive after the contract decision, not before it. Until then, the public has to rely on observational figures, trust statements and a supplier’s own case studies. That is a thin basis for a £330m decision, even if the figures are accurate.
Who is affected by the contract decision?
Patients are the most obvious group affected. The FDP is meant to shorten waits, speed discharges and help cancer teams track pathways. If the tools are not used, those gains may not reach patients at the trusts that signed up. If they are used, the benefits may be real but unevenly distributed. The published data suggests both patterns are possible.
Clinicians and trust staff carry the adoption burden. Some trusts told The Guardian the platform did not add enough functionality to justify the work. University Hospitals of North Midlands said it did not intend to roll the FDP out. Other trusts reported benefits, which suggests local conditions matter. A platform that works in one hospital may duplicate systems in another.
Taxpayers fund the contract. The £330m figure covers up to seven years, and the government must decide whether to extend it. Two Commons committees have recommended against extension. The health select committee has called for ministers to pull the deal. Chi Onwurah, who chairs the science and technology committee, asked whether “the hype surrounding Palantir’s products matches reality”.
Palantir is fighting back. Its UK and Europe chief executive, Louis Mosley, said ending the contract would be “putting politics above patients”. He accused critics of “Palantir Derangement Syndrome”. The company hired Lord Tom Watson, the former Labour deputy leader, as a full-time senior lobbyist in October 2026. It has also written to critical MPs and published videos defending its record.
The government is under pressure from several directions. Health secretary Yvette Cooper faces calls to use the break clause. NHS England wants to avoid disruption to a platform that some trusts rely on. The National Data Guardian, Nicola Byrne, warned in October 2026 about supplier decisions. Public and professional trust should carry “substantial weight”, she said.
Her letter added a different concern. She said a significant number of local organisations were reluctant to adopt the FDP. Reasons included concerns about the supplier, the overhead of managing objections and a preference for other solutions. She warned that rising National Data Opt-Out rates could make national health data less complete and representative.
What are the limits of the FOI findings?
The FOI findings have limits. Medact is a campaign group that wants Palantir out of the NHS. The Guardian says the findings emerged from more than 100 FOI requests, but the article does not publish a full sampling method. The 29 trusts are a subset of the 137 on the benefits list, not a random sample of every NHS organisation.
The replies also use different definitions. Some trusts said the platform was not live. Others said they were not using its applications. Some may use a small part of the platform for cost calculations, as NHS England notes. The replies were given at different times, while NHS England’s list is a quarterly snapshot. A trust that was dormant in July may have adopted a tool by September.
None of this proves that the FDP has no value. It does show that the public benefits claim is softer than it sounds. A trust can be “realising benefits” without transforming clinical care. The OSR has already asked for clearer labels. The Guardian’s investigation suggests the labels still confuse the organisations themselves.
The dispute is also a test of how public bodies describe AI projects. Governments often publish adoption counts because they are easy to measure. They are less comfortable publishing evidence that a tool is unused. The FDP is not the first public IT programme to face this problem. It is an unusually well-documented example because the FOI responses are public.
What would stronger evidence look like? A controlled comparison, a clear definition of use and trust-level data published before the contract decision. NHS England has promised more trust-level transparency. The Imperial College evaluation could provide some of that, but its timing limits its influence on the March 2027 decision.
What happens next?
The immediate decision point is December 2026. NHS England’s chief data officer, Ayub Bhayat, gave evidence to MPs in June 2026. He said the supplier would need to be told by December 2026 if the contract was not being extended. The initial three-year term ends in March 2027. That date is the break clause that campaigners want the government to use.
A successor is unlikely to be ready by March 2027. Ministers have said normal procurement timelines make that difficult. Ending the contract could therefore cause disruption, even if the government wants a different supplier. That trade-off is central to the decision. It is not a simple choice between Palantir and no platform.
Several reviews are still to come. Imperial College’s evaluation is expected then. The National Audit Office will publish its findings in winter 2026/27. The OSR will continue to monitor how FDP figures are used. The Guardian says the regulator will take the new “realising benefits” complaint forward under its casework function.
Palantir’s lobbying will continue. The company has an incentive to keep the contract and a platform to defend. NHS England has an incentive to show that its rollout is working. Trusts have an incentive to avoid disruption. Patients have an interest in whether the tools actually shorten waits and improve care.
If the government extends the contract, it will need to explain why the benefits claims are reliable despite the FOI findings. If it does not, it will need to explain how services will continue without disruption. Either way, the December update and the March deadline will force the issue into the open.
The next quarterly update is due in December 2026. Watch whether NHS England changes the labels on its benefits list. Watch whether the 29 trusts remain on it. Watch whether the government publishes the evidence behind its decision before it extends the contract. Those details will matter more than the headline adoption numbers.