Twenty-nine NHS hospital trusts listed by NHS England as benefiting from its Palantir-led data platform told investigators that it was not operational or that they were not using its applications, according to Freedom of Information responses reported by the Guardian. The findings raise questions about how use and benefits are counted ahead of a decision on whether to extend the contract.
The Guardian’s investigation, published on 2 November 2026, was based on more than 100 FOI requests obtained by Medact, a healthcare campaigning organisation, and shared with the newspaper. Fish 4 News has reviewed the reporting alongside NHS England’s own published figures, official data-protection explanations and parliamentary briefings. We have not independently obtained or audited the underlying FOI response set.
What the Freedom of Information responses show
The 29 trusts are not simply organisations that have declined to sign up. They appear on an NHS England list of hospitals said to be realising benefits from the Federated Data Platform (FDP), yet their FOI replies described no live system, no use of FDP applications or no active operational use.
Examples reported by the Guardian include Epsom and St Helier, which said it was not using the FDP at that time; Kent and Medway, which said it was not yet live; and Mersey Care, which also said it had not gone live. Nottingham University Hospitals said no implementation timetable had been agreed. University Hospitals of North Midlands said it had no intention of rolling the platform out.
There is an important qualification. A trust may have technical infrastructure installed or use a limited submission or costing function without using applications that assist with waiting lists, theatre scheduling or discharge decisions. Torbay and South Devon, for example, told the Guardian that infrastructure was in place to support data submission but that it was not operationally using the FDP. That distinction matters when interpreting the headline number.
Connected, live and benefiting are not necessarily the same
The NHS England uptake and benefits page described 139 trusts as live and 137 as reporting benefits in its May 2026 snapshot. The House of Commons Library briefing, dated 29 October 2026, cites later August figures of 142 of 202 trusts in live operational use and 170 signed up. These are dated snapshots, not figures that should automatically be treated as directly comparable with each trust’s FOI response.
NHS England’s definition of benefit can include back-office analysis and data submissions, not just clinical applications. It is therefore possible for a trust to be counted under the national programme while reporting that it is not using the patient-care tools. What the FOI findings challenge is how clearly that distinction is communicated to Parliament and the public.
Claims about clinical impact also require care. NHS England and Palantir point to more operations, shorter discharge delays and improved diagnostic timeliness at organisations using particular tools. But NHS England’s own methodology acknowledges that before-and-after comparisons cannot prove that FDP caused those changes because other factors were not controlled for.
In July 2026, the Office for Statistics Regulation welcomed stronger caveats but said the way results had been communicated needed improvement. The regulator told the Guardian on 2 November that it would examine the latest concerns under its casework process. A commissioned independent evaluation by Imperial College was, according to the Commons Library, expected to be published after the contract-extension decision.
What the technology does with NHS patient information
The FDP is a software layer designed to connect information held in existing NHS systems, allowing authorised teams to analyse waiting lists, plan surgery, coordinate discharges and perform administrative reporting. It is not, according to the Commons Library, a single central database holding every patient’s complete medical record. Participating organisations have their own instances and remain responsible for decisions about the data they control.
Under NHS England’s privacy notice and contract explainer, Palantir acts as a data processor following NHS organisations’ instructions. NHS England says the contract bars the company from selling, marketing or otherwise commercialising patient information, including using it to train Palantir’s own AI models. It says data is stored and processed in the UK and governed by access controls, security vetting and audit logs.
Those restrictions should not be confused with a claim that no contractor can ever access identifiable data. In its clarification to the National Data Guardian, NHS England said three Palantir engineers had administrative access to the National Data Integration Tenant (NDIT), a specific part of the programme. As at 15 June, a further 33 engineers from various suppliers had more limited, task-specific access. NHS England acknowledged that authorised engineers could, for technical support purposes, access identifiable information within that environment, although it says such access is controlled and patient records are not routinely viewed.
The National Data Guardian confirmed in July that an earlier Data Protection Impact Assessment had described those access arrangements inaccurately. NHS England accepted that error and promised corrections. The Guardian said the National Data Guardian could not independently verify NHS England’s assessment that the access was technically necessary.
None of these findings establishes that Palantir sold patient records, used them to train its own models or accessed them unlawfully. The documented issue is whether technical access arrangements and their public explanation have been sufficiently clear and accountable.
What NHS England and Palantir say
Responding to the Guardian, NHS England maintained that nearly 140 trusts had reported benefits from the platform’s range of tools. It said these helped staff join up care and work more efficiently. Its published figures expressly include back-office products, meaning a reported benefit does not always mean an NHS hospital is running one of the clinical applications.
Palantir told the Guardian that its software was helping trusts deliver better care, pointing to about 117,000 additional operations, a 14.3% reduction in delays affecting long-stay patients and a 5.6% improvement in the proportion receiving a cancer diagnosis or an all-clear within 28 days. Those are the company’s cited outcomes, not independently established proof of effects caused by the FDP.
Palantir’s UK and Europe chief executive, Louis Mosley, has argued that ending the contract would put politics ahead of patients. The company rejects criticism of its involvement and has defended the procurement and the practical benefits of its technology. Critics include campaigners concerned about privacy and the company’s other government contracts, but those objections are distinct from evidence about the NHS programme’s actual performance.
Why the next contract decision matters
The contract, awarded in November 2023 to a consortium led by Palantir Technologies UK, is valued at up to £330 million over seven years. That figure describes the potential contract value, not proof that £330 million has already been spent. The initial commitment runs from March 2024 to March 2027, with a contractual option not to extend. According to the House of Commons Library, the supplier must be informed by December 2026 if the initial term is not to be extended.
Two parliamentary select committees have recommended against extending the agreement, while government officials have warned that replacing an embedded national system by March 2027 could prove difficult or disrupt services. NHS England says contractual safeguards, NHS-owned components and exit planning reduce the risk of being locked into one supplier. The question is therefore not only whether the project is worth continuing but how a change would affect organisations already using it.
What remains unclear
- How many trusts use FDP patient-care applications regularly, compared with technical connectivity or reporting-only functions.
- How NHS England reconciles each trust’s local FOI description with the category of benefit shown in the national list.
- What measurable clinical or financial benefits can be attributed specifically to the platform rather than wider NHS changes.
- The total committed and spent to date, the costs of changing supplier and the practical effect on hospitals that rely on FDP tools.
- Whether the government will exercise the March 2027 non-extension option and what it will publish to explain the decision.
The National Data Guardian’s October advice supported better use of health data but urged ministers to place substantial weight on public trust, evidence, interoperability and data sovereignty. These concerns do not by themselves show patient data misuse, just as claims of improved performance do not by themselves establish causation. Both require verifiable evidence.
Sources
- The Guardian, 2 November 2026, FOI investigation
- NHS England, FDP uptake and benefits figures and methodology
- House of Commons Library, FDP briefing, 29 October 2026
- NHS England, privacy notice and contract explainer
- NHS England, NDIT supplier-access clarification and National Data Guardian’s July response
- Office for Statistics Regulation, July 2026 casework statement
- National Data Guardian’s October 2026 advice to ministers
