Contents
2. Non-adoption, delay and roll-back
3. Quality concerns, other software and alternative local solutions
4. Additional costs related to the rollout, with some Trusts receiving additional money from DHSC
5. Governance concerns and unclear accountability regarding AI discharge summaries
6. Benefit claims rely on weak attribution
7. Increasing data security concerns
9. Public trust and staff buy-in are insurmountably low
Introduction
This report is prepared by Medact to summarise our current, updated knowledge of issues with the rollout of the Federated Data Platform. Our previous report on this issue is available here in the Medact Briefing: Concerns Regarding Palantir Technologies in NHS data systems. This report focuses on issues with the up take and evaluation of the roll-out.
The majority of information in this document is from two rounds of Freedom of Information (FOI) requests that were sent to every trust in England. The aim of these FOIs was to create an accurate picture of the rollout of the FDP. A community-managed public database of these FOIs is available here: https://fdp-rollout.onrender.com/.
To our knowledge, this is the only publicly available in-depth assessment of the rollout at the trust level. Where possible, we have compared this to the NHS England list of organisations ‘live’ with the FDP and ‘realising benefits’. The main limitation of this database is that FOIs may be up to 6 months out of date, depending on when reply was received, and a trust’s rollout stage may have changed in the interim. Where we have referenced an FOI, the relevant correspondence is linked and we have included the date. A third round of FOIs is currently underway, and we would be happy to update this document once these replies are analysed.
Where information is obtained from other sources, such as public documents or articles, this is stated and hyperlinked.
1. Unclear adoption metrics
NHS England (NHSE) reporting seems to be using delivery metrics, such as if organisations are ‘signed up’ or ‘live’, as a proxy for success or ‘realising benefits’. We find this concerning as this is a measure of delivery, not benefit, especially in a context where many trusts believe that signing up to the FDP is mandatory (see below). It is also notable that the National Infrastructure and Service Transformation Authority Green Rating for the programme also relies on delivery metrics.
In our briefing (p.30), we highlighted that signing an MOU with NHSE does not equate to implementation of the FDP. In response to concerns raised by South Warwickshire’ Trust regarding signing the MOU, NHSE confirmed that doing so didn’t commit them to any timeframe for adoption.
We are concerned that signing an MOU, ‘live’ and ‘realising benefits’ are being conflated in NHSE reporting. For example, in February 2025, NHSE claimed that 96 trusts (just over 40%) had ‘signed up to’ the FDP across England. However, NHSE later admitted in a FOI that only 34 trusts (just under 15%) were actively using the platform and its products (called ‘instances’), while the other 62 trusts had only ‘signalled their intent’ to do so.
The FOIs we conducted highlight that having an active FDP instance or being ‘signed up’ to using the FDP does not automatically equate to actual usage or benefits. For example:
- East Suffolk and North Essex NHS Foundation is listed by NHSE as ‘live’ with the FDP and ‘realising benefits’. However, the trust responded to an FOI in February 2026 stating: “The Trust has an FDP instance; however, it is not actively in use due to a change to a new EPR system on 2nd October 2025. There are currently no plans to re-activate our FDP data flows using our new EPR”.
- Cambridge University Hospitals Foundation Trust is listed by NHSE as ‘live’ with the FDP. However, in January 2026 in an FOI response, the trust stated: “the Trust has an FDP instance. We do not have any FDP products in use at this time.” The trust stated it was planning to use Optica, with no specific implementation date.
- The Royal Devon University Healthcare NHS Foundation Trust is listed by NHSE as ‘live’ with the FDP and ‘realising benefits’. However, the FOI response in Jan 2026 stated: “A local instance is in place. No products are currently live, but the Trust is planning to roll out the inpatients/theatres product within the next six months. The FDP is also used for a small number of national data submissions”. It is unclear how they are already listed as ‘realising benefits’ or what metrics are used to assess that, given, if they did go ahead with roll out, it would have been very recent.
Some trusts listed as ‘live’ and ‘realising benefits’ may only have limited usage. For example:
- Northern Lincolnshire and Goole Hospitals NHS Foundation Trust is listed by NHSE as ‘live’ with the FDP and ‘realising benefits’. However, an FOI response from January 2026 stated it is only using Optica, and that: “Other modules have been assessed but not deployed due to being limited compared to locally available solutions.” It also stated that: “the Trust is not planning to migrate existing data infrastructure onto FDP at this stage, which is where additional costs would likely arise, however, it still requires data infrastructure to operate in parallel to FDP. In addition, the Trust is part of a provider group which the FDP does not natively support the consolidation of multiple organisations under a single instance”.
- Milton Keynes NHS Foundation Trust, which was one of the initial FDP pilot sites, stated in an FOI in December 2025 that its two FDP apps are only in “some use”.
- Bradford Teaching Hospitals NHS Foundation Trust said in a January 2026 FOI that: “the OPTICA Discharge tool is currently being tested. There are no plans in the pipeline for wider use of the FDP or other products”.
- York and Scarborough Teaching Hospitals NHS Foundation indicated in a February 2026 FOI that it is only using the OPTICA app, stating: “there are no firm plans for any other FDP apps or products”.
- Nottinghamshire healthcare NHS Foundation Trust is listed by NHSE as ‘live’ with the FDP, but a February 2026 FOI stated that the trust was: “not yet fully deploying”.
- With regards to OPTICA, we understand that this FDP functionality has multiple similar equivalents that already exist within NHS data systems.
- For example, in an April 2026 FOI correspondence, Worcester Acute Hospitals NHS Trust stated it did not implement OPTICA because: “we were successful for constitutional standards funding for a bed management system that integrated with our electronic patient record, so OPTICA was not required.”
There is also inconsistency in reporting of usage of particular apps. For example:
- In June 2026, NHSE reported that North Tees and Hartlepool NHS Foundation Trust implemented the Care Coordination Solution (CCS) FDP app on 01/06/2023. However, in February 2026, the trust reported that the only app it uses is Optica.
- After conversations with staff at some hospitals who are using FDP apps, we are also concerned that FDP apps may be ‘in use’ but not at all used in practice for various reasons. This highlights that adoption, and even on paper ‘usage’ of an app, does not automatically equate to clinical use, impact, or patient benefit.
Pressure on trusts and ICBs to adopt
The inconsistent adoption metrics and lack of consistency between MoU signing and actual use of the FDP may reflect a level of pressure felt by trusts to adopt the FDP.
For example, in response to FOIs, multiple trusts replied believing that there was a mandate to implement the FDP:
- Humber Teaching NHS Foundation Trust replied to an FOI in December 2025 stating that: “the trust is not using the FDP, current medium planning guidance mandates implementation by the end of 2028/29”.
- Homerton Healthcare NHS Foundation Trust replied to an FOI in February 2026 saying that: “While implementation of the FDP is not formally mandated, trusts are being asked and encouraged to use it. As the FDP is being used by other trusts, the Trust does need to engage with it in that context. If NHSE determines that our performance will be judged based on use of the FDP, then the Trust will need to consider usage. It is worth noting that the FDP is free to NHS organisations, where other solutions are not”.
- University Hospitals Bristol and Weston NHS Foundation Trust replied to an FOI in February 2026 stating that: “The use of the FDP across the NHS is part of nationally mandated planning expectations, with wider adoption required by NHS England over the coming years”.
- North Bristol NHS Trust replied to an FOI in January 2026 stating: “Some limited data flows are mandated nationally by NHS England, and our current participation focuses on finance and performance products that do not require identifiable patient data. […] The use of the FDP across the NHS is part of nationally mandated planning expectations, with wider adoption required by NHS England over the coming years. These are national policy decisions, rather than choices made locally by individual Trusts.”
- Tameside and Glossop Integrated Care NHS Foundation Trust stated in an FOI in January 2026 that it was using the FDP for “acute data sets as per NHS England requirements.”
Login data shows shallow and low uptake
A third of trusts with access to Palantir’s Federated Data Platform (FDP) seem not to have logged on to the apps for the past year, the BMJ has reported.
Data obtained under freedom of information (FOI) laws contrast with rollout claims that NHS England, Palantir and ministers have repeatedly used to justify the controversial software.
Officially, 139 NHS trusts are “live” with the FDP. This number has been repeatedly cited by officials as an indicator of its success.
Rob Thompson, for example, the chief digital, data, and technology officer across the Department of Health and Social Care and NHS England, told the Commons’ Health and Social Care Committee last month that this metric indicated the FDP’s popularity among trusts. “I look at the number of 139 acute trusts having signed up to use it as an indicator of whether they want or need it, and that for me is the benchmark,” he said. But new data indicate that a major portion of trusts are not embracing the rollout.
Under FOI rules, NHS England released internal data showing that only 85 trusts recorded any user activity across the eight FDP apps that NHS England had data for in the year up to June 2026. These apps were, specifically, Cancer 360, Crisis Response, Inpatient CCS, OPTICA, Outpatient CCS, Patient Led Validation, RTT Validation and Shared PTL.
There seems to be no record of the other 54 trusts logging into these apps even once, and a further 46 used just one of the eight apps. NHS England’s data also indicate that, even among 85 trusts that have logged in to these FDP apps, actual usage is often low.
2. Non-adoption, delay and roll-back
Multiple trusts replied to FOIs stating that they did not intend to adopt the FDP, or between rounds of FOIs had delayed adoption of the FDP. Whilst the reasons for this are unclear, it could be due to concerns regarding the quality of the platform (see section 3), staff and patient resistance (see section 9), a focus on other digital priorities, or other barriers.
The latest minutes from the FDP check and challenge group, published in April 2026, highlighted some potential reasons for slower uptake in some trusts. The Director of Data Management and Transformation noted that variation was largely driven by differing local priorities and pressures, including competing programmes such as EPR implementations, as well as the presence of existing local solutions. The group noted that, alongside operational factors, some organisations may also hold concerns relating to the FDP supplier which could influence perceptions, trust and engagement, and this was recognised as a consideration within wider stakeholder sentiment.
For example, multiple trusts have declined or delayed adoption of the FDP:
- In January 2026, Central and North West London NHS Foundation Trust responded to an FOI stating: “No FDP instances, applications or products are currently in use or planned for rollout at this time.”
- University College London Hospitals replied to an FOI in January 2026 stating: “UCLH is not using the FDP yet. We do not have a specific rollout date planned”.
- In May 2025, Barking, Havering and Redbridge University Hospitals NHS Trust replied to an FOI stating: “the infrastructure for the connecting of the FDP is already established” and they were “assessing multiple applications”. However, in January 2026, it replied to another FOI stating: “BHRUT is not using the Federated Data Platform and does not currently plan to implement it locally. As such there is no rollout date, and no onboarding or implementation activity is planned.”
- In January 2026, Sheffield Teaching Hospitals NHS Foundation Trust replied to an FOI indicating that it is not using the FDP and has “been in discussion with the NHSE FDP team to understand more about the platform and its potential use”.
- Kings College Hospital replied to an FOI in December 2024 that it was looking into the FDP actively. It had just implemented a new EPR system by EPIC but was positive about the FDP being rolled out by Q1 25/26. However, in January 26, the hospital said in a further FOI response that: “Kings is not currently using Palantir, although as a provider of NHS services, we may use the system going forward to support the sharing of patient information between providers, which is designed to improve the care provided by health services”.
- Princess Alexandra Hospital NHS Trust stated in an FOI in February 2026 that it had “no current roll out date and not using FDP”, despite initially saying in a May 2025 FOI that it would commence engagement.
In February 2026, Health Service Journal reported on some of these findings in a piece titled “Flagship trusts still not using federated data platform”. It noted that acute trusts that were not using the platform included: “University Hospitals Birmingham Foundation Trust, Guy’s and St Thomas’ FT, the Northern Care Alliance FT, Nottingham University Hospitals Trust, the Royal Free London group, Sheffield Teaching Hospitals FT, University College London Hospitals FT, and Frimley Health FT.” HSJ also noted that: “a number of smaller but highly prestigious specialist acute trusts are also not using the FDP. They include leading cancer specialist the Royal Marsden FT and the internationally renowned Great Ormond Street Hospital for Children FT.”
In addition to the above, some trusts seem to have rolled-back use of the FDP:
- Cheshire and Wirral Partnership NHS Foundation Trust developed a mental health crisis care app, in collaboration with Palantir, that won multiple awards. However, an FOI response from May 2026 confirmed that this app was no longer in use, and it was not using any other apps at this time.
- University Hospitals of Derby and Burton NHS Foundation Trust replied to an FOI in 2025 stating that it was using four FDP products including the referral to treatment (RTT) validation tool. In its January 2026 FOI reply, the trust stated it was using only three, with no mention of RTT.
3. Quality concerns, other software and alternative local solutions
In our March 2026 briefing, we outlined quality concerns regarding the FDP, particularly that local and regional data architecture in some places already exceeds the capabilities being offered by the FDP (p.27), which was also reported by Democracy for Sale in 2025.
The briefing also contains concerns regarding the risk of vendor-lock in, and interviews with NHS data analysts who raised concerns that the software depends heavily on Palantir staff for changes, and they have limited autonomy over the analytical capability.
For example:
- A 2024 FOI reply from Leeds Teaching Hospitals NHS Trust stated that: “From the descriptions we have of these FDP products we believe we would lose functionality rather than gain it by adopting them”.
- In February 2025, the NHS Chief Data and Analytical Officer Network (CDAON) wrote an open letter to NHS England Chief Digital and Information Officer, Ming Tang, detailing objections to the FDP’s rollout, including: “we already have similar tools in use that presently exceed the capability and application of what the FDP is currently trying to develop or roll out at a system level”.
- In early 2025, Greater Manchester (GM) ICB reported that: “[Palantir’s platform] does not currently have any system-level products that offer the same or better functionality, compared to the custom-built system already in use for NHS GM”.
- An investigation by Corporate Watch in 2025 interviewed NHS workers, some of whom cited quality concerns regarding the platform, with one data analyst stating: “Not only could similar functionality have been delivered at a fraction of the cost, but the existing tools are already better integrated, more intuitive, and more conducive to collaboration”. As well as public trust concerns, NHS data analysts described frustration and annoyance. One said: “there is widespread dismay among data professionals within the NHS, many of whom feel they are being silently forced to adopt this platform”. This report contains a significant number of quotes and concerns from NHS analysts and engineers and we would recommend reviewing it separately.
In recent FOIs, some trusts highlighted their preference for locally available solutions. For example, Northern Lincolnshire and Goole Hospitals NHS Foundation Trust is listed by NHSE as ‘live’ with the FDP and ‘realising benefits’. However, an FOI response from January 2026 stated it is only using Optica, and that: “Other modules have been assessed but not deployed due to being limited compared to locally available solutions.”
Financial Times published an article in April 2026 which included sceptical comments about the FDP from multiple senior NHS data leaders and officials. Direct quotes from senior NHS officials taken from the article include:
- A senior NHS official, who runs a major hospital’s data system, said: “I’ve had a good look at what it does and it’s all pixie dust and rainbows. We’ve already got access to much richer data sources than this.”
- “The biggest issue is that this is an expensive tool being used to do things that could be achieved much more cheaply and easily at a local level,” noted another health official “For £330mn, what we’ve got is essentially a database with some applications on top”.
A document analysed by Democracy for Sale in May 2026 revealed that the speed and processing time of the FDP is substantially worse than pre-existing NHS software, evidenced by a briefing seen by Democracy for Sale prepared for NHS England’s senior leadership team for data and analytics. The article states that “Processing time is reported as taking 4-5 minutes compared to 30 seconds (in the current system)”, even after “optimisation” changes. This translates to the FDP being 8-10 times slower in these cases than existing software. A senior NHS source told Democracy For Sale that the runtime is even slower than reported – “I’d be happy with four to five minutes, but I’m waiting closer to ten”.
In April 2026, a group called NHS Analysts Together, representing data engineers, analysts and other digital workers within national and regional NHS bodies launched an open letter advocating for the use of the break clause, as well as “The ‘development and implementation of an NHS Ethical Procurement Framework” and “A consultation with NHS data and digital professionals regarding the technical merits of existing NHS data infrastructure that could be learned from, scaled or replicated.”
The letter notes a number of concerns, including that:
- “Palantir’s FDP is regressive in many organisational contexts. The narrative that the FDP represents a technological step forward for NHS data infrastructure does not reflect the reality experienced by many of our organisations, and we note with concern the push for FDP products above other data architecture options.”
- “A number of NHS bodies have, over many years and with significant investment of professional expertise, developed sophisticated local and regional data platforms – including linked dataset environments, population health tools, and advanced analytics capabilities – that in several respects exceed the functionality currently offered by the FDP.”
- “For these organisations, adoption of the FDP does not represent progress. It represents regression – the replacement of mature, fit-for-purpose, NHS-managed infrastructure with a commercially owned platform, which introduces significant new risks around data sovereignty and commercial dependency. We are concerned that this reality has not been adequately represented in the business case for the FDP, and that the platform has been presented as universally superior, when the evidence from practitioners on the ground does not support this.”
- “We are concerned by the significant local opportunity cost that investment in the FDP comes with, both at the software and professional level. With Palantir maintaining significant control over the core platform and the intellectual property rights of its associated products, NHS institutions risk depending on this software and Palantir staff for the maintenance and design of their data infrastructure. We do not believe this significant vendor lock-in is a sustainable option for NHS digital infrastructure”.
- “At the same time, many of our colleagues, including experienced data and digital professionals who manage essential workflows and could manage alternative data architecture to the FDP, are being forced into redundancy. Data analysis expertise in a healthcare context is not a commodity function that can be automated away. It requires clinical knowledge, organisational understanding, contextual judgement and professional accountability. These are professional capabilities that no platform, however sophisticated, can replicate.”
A London Digital Transformation Portfolio Board paper from February 2025, obtained via FOI, identified that risks of the FDP in North West London included: data latency affecting adoption and use of the tools; lack of write-back functionality reduces usage and benefits; lack of write-back creates lost actions and potential clinical harm; and, incomplete recording of activity completion understates usage.
An article published on 10 April 2026 in Health Services Journal reported that trusts using EPIC have largely rejected the FDP, highlighting that the FDP is not the only software able to meet local needs. EPIC is an electronic patient record system, which provides extensive features and functionality. Almost all Trusts that adopted it (eight out of nine), have rejected or minimised the use of FDP.
The assessment of the FDP programme must not only example the alleged benefits from the FDP, but also the significant opportunity cost of choosing an ‘FDP first’ policy over investment in in-house or other locally-led innovation. Trusts may report benefits from the FDP, especially trusts that have struggled to finance investment in their own digital systems, but this does not mean that Palantir software is the only solution for those trusts.
4. Additional costs related to the rollout, with some Trusts receiving additional money from DHSC
Significant costs related to the programme at the national and local level reflect not only a need to examine the value for money of this contract, but also the significant opportunity costs to local and in-house innovation.
In private conversations, the offer of the FDP as ‘free’ has been cited as a major motivating factor for trusts to adopt the it. This is unlikely to continue indefinitely, especially once the initial contract term is up, and also does not reflect other local implementation costs. The lack of budget for trusts themselves to invest in their own digital capacity, and the need to rely on ‘free’ software from NHSE, is in itself concerning.
Overall, DHSC has reported that the budget for the whole life of the FDP will be just over £1 billion. The department’s own Cost Narrative said that “compared to financial year 23/24-Q4, the project’s departmentally agreed Whole Life Cost at 24/25-Q4 (measured in £m) remained at 1042.”
DHSC documents (here and here) show that some trusts have received millions of points of additional funding to rollout the FDP, including ones which are then used as examples of national success of the programme:
- Between 2023 and 2025, University Hospitals Leicester received £2.75 million from DHSC to facilitate “the procurement of the federated data platform”. The trust, however, did not declare this in FOIs about the cost of the rollout. In response to FOIs regarding extra costs related to the FDP rollout at the trust, University Hospitals of Leicester and University Hospitals of Northamptonshire Group replied that: “the Trust does not have a Federated Data Platform (FDP) budget. The Trust has not had, and do not expect any costs that are not provided for by the national funding”. Neither trust mentioned the extra funding University Hospitals Leicester received in assistance for the rollout across the group.
- Between 2023 and 2025, Chelsea and Westminster received £3 million from DHSC for the rollout of the FDP.
- In total, DHSC gave out £6.75 million in 2023-2024 and £7.7 million in 2024-2025 to individual trusts for “facilitating the procurement of the federated data platforms, the privacy enhancing technology and the products and services that enable the improvement of services”.
The Freedom of Information requests revealed that there are costs associated with the adoption of the FDP at the local level that are not covered by the national NHSE contract. Examples of this from the FOIs include:
- St George’s University Hospitals NHS Foundation Trust responded to an FOI in January 2026 stating that it was not using the FDP and “The Trust expects significant additional costs related to staff and infrastructure in the future however these costs have not been locally budgeted yet”.
- In January 2026, Barnsley Hospital NHS Foundation Trust confirmed in an FOI that it spent £22,000 on the FDP rollout at the local level. This includes: “10K technical resource cost approximately. 2K training staff costs. 10K Technical infrastructure costs virtual server and networking”.
- Imperial College Healthcare NHS Trust responded to an FOI in September 2025 saying that: “costs covered directly by the Trust related to the FDP in 2024/25 were £452,653. The Trust did not incur any direct costs in previous years”. The trust detailed that “this funds a digital operations team – one of their main objectives is to support the deployment of the FDP and maximise its benefits. This covers everything from staff training to technical/data engineering support, project management and implementation expertise. The team also has a wider role supporting the implementation of other digital tools in the Trust, such as reporting dashboards”.
- Croydon Health Services NHS Foundation Trust responded to an FOI in February 2026 stating that: “Trust resource was required to co-ordinate training sessions, set up data feeds and validation of data prior to go-live of the app. This was not funded by the FDP”.
- The Shrewsbury and Telford Hospital NHS Trust stated in an FOI response from February 2026: “it has relied on external contractors who work closely with internal trust teams and the national FDP team, on an interim basis for the duration of the project”; and “Costs related to external contractors are covered by the Digital Services capital allocation for project implementation”.
- Other trusts reported absorbing costs related to project management, operational delivery, technical set up, staff time, project implementation within existing teams and budgets. This included University Hospitals Coventry and Warwickshire NHS Trust, Cheshire and Wirral Partnership NHS Foundation Trust, Guy’s and St Thomas’ Foundation Trust, Newcastle Hospitals NHS Foundation Trust, and South Tees Hospitals NHS Foundation Trust.
5. Governance concerns and unclear accountability regarding AI discharge summaries
A pilot of AI-assisted discharge summaries is currently underway at Chelsea and Westminster Hospital NHS Foundation Trust, which has prompted concerns over governance and accountability of AI within the. We argue this also represents significant scope creep beyond the use cases of the FDP.
An article published 15 April 2026 in Health Services Journal titled “The Palantir PR Project” raised concerns about this, which neither NHS England nor Chelsea and Westminster Hospital NHS Foundation Trust have responded to. In particular, questions have been raised about the regulatory framework for the AI discharge summary tool. FOI requests revealed that it was “developed and piloted under MHRA’s in-house exemption”, which meant it did not require MHRA registration or UK Conformity Assessed (UKCA) marking. However, an individual who worked on the project has said that the tool itself was primarily developed by Palantir engineers, with less involvement from the trust – something that might not be covered by the exemption. Additionally, serious concerns around patient consent remain unanswered, as a total of 27 patient discharge summaries were generated by the AI tool using live patient data. ‘Implied consent’ could have been used if it was providing direct care. However, if that was the case, there is a suggestion that it would need to be classed as a medical device – and ‘implied consent’ would not apply.
Another area of criticism to the AI discharge summary tool has been the fact it is built on-top of pretrained large languages models (LLMs). LLMs are highly stochastic which makes testing and safeguarding robustly against bias extremely difficult. It is not clear what steps have been taken to mitigate the risks of bias in this tool, which is surprising given the impact the contents of a discharge summary can have on the care a patient subsequently receives. It is of extreme importance that AI-based tools like the discharge summary tool do not in any way entrench pre-existing inequalities in healthcare. The well known tendency for LLMs to hallucinate and invent new information also means that any AI-generated outputs need to be carefully checked by clinicians for errors, which potentially mitigates any time-saving benefits the tool could provide. In its FOI response in February 2026, Chelsea and Westminster Hospital NHS Foundation Trust conceded that the pilot “was at too early a stage to draw meaningful conclusions” and that “the dataset is insufficient for robust statistical analysis or generalised findings”. We are not aware of any developments in bias-monitoring since then.
The latest minutes from the NHS England Data Transformation Check and Challenge Group, published in April 2026, highlight further concerns: “It was confirmed that the Data Protection Impact Assessment (DPIA) for the AI Assisted Discharge Summary remained on track but could not be finalised until the product reaches pilot stage”.
In these minutes, published on June 1st, a member from the British Medical Association challenged whether the current classification of AI-Assisted Discharge Summarisation (AADS) remained appropriate as AI medical device regulation continues to evolve. The NHS England FDP Programme Team acknowledged that the regulatory landscape was evolving but provided assurance that AADS remained compliant with current legislation and regulatory advice.
6. Benefit claims rely on weak attribution
Experts have raised concerns that the benefit claims regarding the FDP are based on weak methodologies and problematic assumptions, such as attributing any benefit seen post FDP implementation to the FDP, without consideration for confounding factors or control comparisons.
NHS England and Palantir have both made significant claims about the benefits provided by the FDP and the scale of the adoption of the FDP. These claims have come under significant scrutiny, and NHS England is currently under investigation by the Office for National Statistics.
Benefit claims have been challenged on a number of grounds, including:
- Benefit claims at specific hospitals may have conflated correlation with causation by attributing all positive change to the FDP, when it may, for example, have been post pandemic recovery. Read more in the BMJ.
- Some national claims rely heavily on a very small number of hospitals. Read more in the Financial Times.
- Some local hospital data used to make national claims was based on potentially faulty data. Read more in the Financial Times.
- The number of trusts meaningfully using the FDP appears to be lower than national rollout figures suggest. Read more in Democracy for Sale and in the BMJ.
In April 2026, the BMJ published a piece critiquing the methodology used to make significant benefit claims at Chelsea and Westminster NHS Foundation Trust, the site of an FDP pilot of the Care Coordination Solution. Firstly, the methodology used to claim that the FDP increased theatre efficiency risked conflating causation with correlation, and other NHS data showed that trusts who were not part of the FDP pilot also showed improvements in theatre utilisation during this period. Secondly, public waiting list data challenged the claim that the FDP pilot resulted in a waiting list reduction. Due to the lack of public methodology the waiting list claim could not be fully examined.
Most other evaluations that underpin the claims about FDP benefits have not been published in full and therefore cannot be scrutinised. However, we have serious concerns regarding the legitimacy of the claim that the FDP has enabled 110,000 additional patients to undergo procedures in theatres. The methodology is described as the following: ‘A cumulative total of additional patients undergoing procedures in theatres by NHS trusts with the Inpatient Care Coordination Solution until end of December 2025, compared to the previous period without NHS Federated Data Platform use.’ This methodology for this figure relies on an enormous assumption that any increase in patient procedures was due to the FDP.
Whilst plans to commission an independent evaluation, we are concerned that this evaluation will be biased towards confirmation, given the tender description of “understanding and demonstrating the impact the programme is having and will have in the future, discern if its objectives have been achieved, capture key learning and demonstrate value for money and accountability to its stakeholders.”
7. Increasing data security concerns
In our March 2026 briefing, we highlighted a number of issues regarding data security within the FDP, including the concerns regarding pseudonymised data. In addition, the data privacy of staff in the NHS must also be considered, given the revelation that Palantir staff are being given nhs.net emails.
In particular, we noted concerns about potential and perceived risk of government abuse of health data in future, and the impact on health inequalities. We are alarmed that since then, it has been reported that Reform UK’s policy document ‘Operation Restoring Justice’ includes a proposal to build a powerful surveillance database using data mined from across UK government, healthcare, financial, and police databases. Palantir UK CEO Louis Mosley then confirmed in the Observer that:
If Reform won the next general election with a “clear public mandate” to allow NHS data to be used for immigration enforcement, then the company would execute the plan.
In March 2026, Health Service Journal reported that:
Almost one-third of NHS trusts using the federated data platform last year were not meeting minimum data security standards. Senior officials considered whether to eject the organisations from the controversial programme, because of “the scale” of the issue. Minutes from a meeting of the FDP data governance group in October, which were published in February, reveal that 25 trusts were not meeting minimum data security and protection toolkit (DSPT) standards.
Following the revelation that some Palantir staff would obtain unrestricted and unlimited access to data within the National Data Integration Tenant (NDIT), which contains patient identifiable information, the website Not With My NHS Data was set up. Over 900 people have made complaints to the National Data Guardian, who issued a statement in response. It stated: “The DPIA we reviewed stated that access to identifiable patient information would be limited to NHS staff with a legitimate need. However, since then, recent media reporting, and subsequent confirmation from the programme team, indicate that some external contractor staff also have access to identifiable patient information within the NDIT environment. We were not aware of this. We have therefore written to the programme to seek clarification on this inconsistency.”
8. Alternative approaches to federating data and improving digital architecture are possible and already happening
Palantir’s FDP is, in short, a combination of different functions provided on one software platform, largely powered by one supplier. None of these functions are uniquely supplied by Palantir, and other providers and/or in-house functionality is available for each of these functions.
Rather than looking for a single-supplier solution to NHS data, we argue that the government should look to invest in sustainable digital architecture based on the principles of open standards, open source, interoperability, data sovereignty, ethics, transparency and co-design based on need.
Through our discussions with experts in the NHS data landscape, we believe that there are a number of viable alternatives to Palantir as a provider of data federation, analytical tools, and operational applications. We argue that an ecosystem approach which harnessed and invested in NHS innovation, coupled with ethical and appropriate involvement of external suppliers where required, as well as sustained investment in data and digital professionals at the national and local level, could deliver genuine improvements to NHS data architecture.
For example:
- Greater Manchester ICB rejected use of the FDP as the ICB created its own warehousing and analytics solution in which data is federated. This allows the ICB to easily view and use data from trusts and other NHS institutions in its remit, while allowing trusts to retain ultimate control over their own data. The solution has successfully garnered local trust and buy-in. It has been designed to be provider-agnostic, with the dual benefit of preventing vendor lock-in and allowing the platform to harness the best providers available to them at any given time.
- In terms of alternatives to the local level operational applications supplied to trusts through the FDP, it has of course been the case for decades that many British SMEs create and sell their own applications to trusts and other parts of the health system. This includes applications for theatre scheduling, discharge planning and other functionality offered by the FDP. Depending on the solution required, these applications do not always require data federation (if the data does not need to leave a trust for example). If federation is required for data, it does not need to be all on the same software or run by the same supplier, especially if interoperability and shared standards are prioritised across the NHS.
- The FDP contains data analytical tools used at the national level. Our discussions with NHS staff indicate that these are difficult to use and give poor user experience, and are outperformed by other existing tools, some of which are open source. Other tools, such as dashboards and visualisations, can be created with an array of software tools and have been for many years.
9. Public trust and staff buy-in are insurmountably low
In our March 2026 briefing, we highlighted a number of issues which may damage public trust and staff buy-in to a Palantir FDP, and damage trust in the NHS as a whole. This includes Palantir’s extensive allegations of involvement in human rights abuses, concerns regarding surveillance and data privacy within NHS data systems, and the lack of public ability to opt-out. Since the publication of the briefing, these concerns have only heightened.
For example:
- Over 81,000 patients have made formal complaints to their local hospitals regarding use of Palantir software.
- After receiving a significant number of complaints, Bristol NHS Group released a statement confirming that it is in an early stage of engagement, and that it has “made NHS England aware of the strength of feeling on this subject”.
- Initiatives to opt-out of data sharing for research and planning purposes at the national and GP level are multiplying, based on the (incorrect) belief that this will opt patients out of Palantir software and the FDP. Please find one of multiple examples here. This confusion and public mistrust is likely to impact a wide range of data sharing initiatives and negatively impact the entire NHS data architecture.
- Greater Manchester ICB further confirmed its decision not to adopt the FDP locally saying it was not going to go ahead with a planned review of the platform, noting in the ICB meeting minutes that: “It was proposed that a paper would be produced in due course to guide a review of the ICB position. […] This paper has not been produced yet and work has not started on this paper because it’s clear that the public concerns have heightened rather than diminished since the deferral decision has been taken and there does not appear to be any compelling evidence that the value proposition for NHS GM from FDP has materially changed in favour of adoption.”
- The UK’s largest union, UNISON, has urged the government to use the break clause in the Palantir contract, citing concerns regarding public trust: “Ministers mustn’t risk damaging public confidence in such a major project that should provide better healthcare for millions of people. Any nervousness among patients about the use of their data could prompt them to opt out and undermine the whole system. It would be far better to make a break from Palantir at the earliest opportunity.”
- The actions of Palantir senior leadership and the company as a whole continue to pose a risk to public trust in the NHS and data systems. For example, Palantir issued a 22 point manifesto on X, proclaiming some cultures’ superiority to others.
Public trust and staff buy-in to the FDP has also been impacted by perceived conflicts of interest and ‘revolving doors’ between Palantir and senior staff promoting the FDP. Examples of this include:
- Senior NHS leaders who have since gone to work at Palantir, highlighted in the Medact briefing.
- Financial Times reported that whilst Matthew Swindells was the Chair of four NHS trusts, including Chelsea and Westminster, he was also an advisor to Palantir through General Counsel, Peter Mandelson’s lobbying firm, which represented Palantir. After the Financial Times article, Swindells resigned from his NHS positions.
- On 8th June 2026, Will Monaghan, Executive Chief Digital Information Officer at University Hospitals Leicester NHS Trust, spoke about the benefits of the FDP at an online event, run by The Kings Fund and sponsored by Palantir. Monaghan declared no conflict of interest in relation to this event. However, Monaghan filed a declaration on 15th May 2026 stating that he was paid £13,000 for “attendance at the Automation Anywhere conference”. The Automation Anywhere conference was financially supported by two ‘platinum sponsors’, one of which is private management consultant Accenture. Accenture is a private company contracted by the NHS to work on Palantir’s FDP consortium.
- In March 2026, Andrew Inchley, chair of the Finance and Investment Committee and member of the corporate trustee board at University Hospitals of Leicester, declared that he is a director of two of Accenture’s subsidiaries. Inchly became a Director at University Hospitals of Leicester in April 2025, after almost two decades working with Accenture.
- Stephen Childs, Palantir UK’s head of health partnerships, was until recently the managing Director of North of England Care Support System (NECS). At NECS, Childs worked with Palantir to develop initial FDP use cases and supported Palantir’s bid for the FDP.
The level of public and staff concern regarding this platform will likely significantly impact and minimise any potential benefit from the programme. The risk of knock-on effects of damaging other NHS data systems, such as GP data sharing agreements and research, is significant.
