Summary

• Foxglove says around 30% of English trusts using Palantir’s Inpatient CCS module are carrying out fewer operations than before adoption.
• The figures challenge broad claims that the Federated Data Platform is increasing surgical activity across hospitals.
• Public-sector technology claims need transparent evidence, not aggregate performance narratives alone.

NHS England and Palantir are facing renewed scrutiny over the Federated Data Platform after freedom of information data suggested that some hospitals using the system are carrying out fewer operations than before adoption.

Campaign group Foxglove says around 30% of English hospital trusts using Palantir’s Inpatient Coordination and Communication Solution, an FDP module intended to support operation scheduling, reported a decrease in total operations compared with the period before they adopted the tool. The figures cited by the group include 13 of 41 trusts using the module that collectively recorded 9,073 fewer operations after implementation.

The findings challenge broad claims that the platform is increasing surgical procedures across the NHS. Foxglove argues that aggregate figures obscure variation between trusts, making it harder for the public, Parliament, and clinicians to judge whether the technology is improving operational performance in practice.

The Palantir contract has been controversial from the outset, largely because of questions around patient data, foreign supplier dependency, and the company’s work with defence and intelligence customers. This dispute moves the argument towards evidence of impact. If a platform is justified partly on its ability to reduce backlogs and improve capacity, performance data has to be detailed enough to test that claim.

The FOI figures do not prove that the system reduced operations. Hospital throughput is affected by staffing, theatre availability, industrial action, funding, patient complexity, discharge capacity, local management, and seasonal demand. A before-and-after comparison cannot isolate the platform’s effect on its own.

The same caution applies to positive claims. If improvements cannot be attributed solely to the FDP, broad aggregate gains need careful explanation. Trust-level variation, other operational changes, baseline capacity, and local implementation conditions all affect whether the technology can be credited with better performance.

Public-sector data platforms often promise better coordination, faster decisions, and more efficient use of resources. Those benefits are difficult to measure once systems enter pressured operational environments. A hospital is not a software demonstration. It is a constrained institution with clinical priorities, workforce pressures, legacy systems, and local processes that can overwhelm a tool’s theoretical capability.

The NHS also has a trust problem around large data programmes. Public confidence depends on safe information use and visible supplier accountability. Controversy over data access can be managed only if the operational case is strong, transparent, and open to scrutiny. Partial, disputed, or difficult-to-interrogate evidence weakens even useful tools.

For Palantir, the NHS contract is a prominent European reference case for public-sector data-platform work. Demonstrable impact across varied hospital settings would strengthen its position. Uneven results, unclear attribution, or limited transparency would give critics a stronger argument that the NHS is taking on supplier and public-trust risk without enough evidence of benefit.

The next stage should be a disciplined evaluation of where the FDP works, where it does not, under what conditions, and at what cost. Public-sector technology earns legitimacy when claims can be tested. On that measure, the NHS and Palantir still have work to do.