Public mistrust surrounding Palantir is becoming a risk to the NHS’s ambitions for wider health-data use, after health innovation minister James Frith warned that controversy around the US technology company could make patients less willing to allow their information to be used for research and planning.

About 60,000 more people used the national data opt-out between mid-May and mid-July, according to figures cited in correspondence between Frith and Commons health committee chair Layla Moran. Frith described the increase as modest and said it would be monitored, while linking the timing to greater public attention around Palantir’s role in the health service.

The figures do not demonstrate that Palantir caused the rise, and the national data opt-out covers secondary uses of confidential patient information across England rather than a single supplier or system. Patients who opt out can still have their records used for direct care, but their confidential information is generally excluded from uses such as research and health-service planning.

Frith nevertheless warned, in correspondence reported by the Guardian, that “It may not be possible to realise the benefits of the 10-year health plan if patients stop sharing their data.” The concern moves the Palantir argument beyond procurement because the quality of health research depends partly on whether enough people continue to accept secondary uses of NHS information.

The platform remains under political pressure

Palantir leads the provision of NHS England’s Federated Data Platform, which is designed to connect information already held across NHS systems so organisations can coordinate care, manage waiting lists, plan resources, and analyse services. NHS England describes the contract as worth up to £330 million over seven years and says the platform can serve as many as 240 trusts and integrated care systems.

The original procurement notice records a contract award value of about £182.2 million excluding VAT, with spending capable of rising towards the larger forecast as services are consumed. NHS England maintains that the health service retains control of data held in the platform and that Palantir is contractually prohibited from commercialising NHS information, using it to train its own AI products, or accessing it for purposes unrelated to providing the service.

Those technical and contractual controls have not ended the political dispute. Critics have challenged Palantir’s work elsewhere in defence, intelligence, immigration, and law enforcement, while supporters point to operational improvements reported by NHS organisations using the platform.

NHS England has itself cautioned against treating reported improvements as proof that the technology caused them, and the statistics watchdog is examining some claims made around FDP outcomes. Waiting-list reductions or discharge improvements can be influenced by staffing, funding, service redesign, management interventions, and changing patient demand as well as the software used to coordinate work.

Trust behaves like infrastructure

The national data opt-out creates a different kind of constraint because the effect appears upstream from any particular analytics system. When patients decide their confidential information should not be used beyond individual care, the available pool for research and planning becomes smaller regardless of whether the eventual analysis is conducted on Palantir technology, another platform, or conventional data infrastructure.

A relatively modest rise does not make an NHS research dataset unusable, although sustained or uneven opt-outs can affect representativeness over time. If particular demographic or patient groups become less willing to participate, researchers may inherit biases that cannot be solved through better computing alone.

Assurances about security therefore have to coexist with confidence in institutions, procurement, suppliers, governance, and the purposes for which data will be used. A technically secure system can still encounter resistance when patients distrust the organisation operating it or believe public bodies have not adequately explained the relationship between a commercial supplier and sensitive information.

The FDP has already provided a case study in that problem. NHS England had to correct its public account of supplier access to identifiable information earlier this summer, reinforcing scrutiny of how the platform’s safeguards are explained as well as how they operate technically.

Frith has also said there is no requirement for NHS organisations to use the FDP and that trusts can prefer alternative systems, after earlier parliamentary language suggested parts of the platform would become mandatory. The Department of Health and Social Care maintains there has been no change in policy and that local organisations have always been able to choose another option where it meets their needs.

That flexibility may reduce immediate concerns around compulsory adoption, but it introduces another implementation question. A national data programme intended to improve coordination and common analysis gains value from widespread adoption, while a fragmented estate of local alternatives can increase interoperability and governance costs even if individual trusts prefer greater supplier choice.

The dispute exposes a constraint that cannot be resolved by contract clauses alone. NHS England can specify who controls data, where it is stored, and what a supplier may do with it, but patients retain their own judgement about whether the wider system deserves confidence. If mistrust begins to reduce participation in research and planning, public sentiment becomes an operational dependency for the health service’s data strategy rather than a communications problem around one technology contract.