There is a particular kind of discomfort that settles over a select committee room when a witness concedes a point his own organisation has spent two years defending. During his appearance before the Health and Social Care Select Committee, Sir Jim Mackey pointed out that an independent evaluation of the Federated Data Platform would be beneficial and necessary. He also stated that he had personally questioned the validity of the figures published in its favour. For a chief executive whose department has repeatedly cited the platform as evidence of digital transformation delivering for patients, this was not a small admission.

In order to connect disparate hospital, general practitioner, and community networks and administer wait lists, bed capacity, and theatre schedules with almost real-time visibility, the health service paid £330 million for the Federated Data Platform, which was developed by the American company Palantir. NHS England has credited it with contributing to more than 100,000 additional surgeries and a 15 per cent reduction in discharge delays. Those figures have underpinned ministerial rhetoric about productivity gains flowing from technology rather than headcount. Mackey's testimony is significant because it reveals a discrepancy between what the center has said and what it is now prepared to fully defend.

The substance of the challenge is not abstract. Critics, including campaign groups scrutinising trust-level data, argue that the bulk of the reported improvement is concentrated in a small number of early-adopting trusts, while a meaningful proportion of participating organisations have seen surgical throughput fall rather than rise. If that pattern holds up under review, the platform's headline statistics look less like a national productivity story and more like a handful of local success cases doing considerable rhetorical work. This distinction is crucial for any secretary who might use the Federated Data Platform as an example for AI-enabled change elsewhere in the health sector, as well as for the Treasury's willingness to fund new digital procurement based on untested benefit propositions.

Mackey's caveat about timing is the more pointed detail. He indicated that any independent review is unlikely to conclude before the contract's break clause point this autumn, when ministers must decide whether to continue, renegotiate or exit the arrangement. That sequencing puts the government in the position of making a multi-year commercial decision before the evidential basis for the platform's benefits has been tested by anyone outside the organisations with an interest in defending it. It is the kind of governance mismatch that tends to surface later in a National Audit Office report, not the kind ministers usually choose voluntarily.

The wider context compounds the difficulty. Internal directives expanding contractor access to identifiable patient data have unsettled clinical staff already wary of the platform, and the British Medical Association has continued to press its objections on the grounds of Palantir's commercial ties to defence and immigration enforcement work in the United States. The Permanent Secretary's recusal from elements of the file over past advisory links to the procurement process adds a further layer of unease about how cleanly this contract was arrived at in the first place, whatever the eventual verdict on its performance.

None of this guarantees the platform will be removed. Ministers have asked officials to prepare contingency plans for its exit, which is a sensible precaution rather than a settled conclusion, and there remains a credible argument that federated data infrastructure, imperfectly executed, is still preferable to the fragmented systems it replaced. But the government's room for manoeuvre has narrowed. It can no longer treat the Federated Data Platform's benefit case as settled while simultaneously commissioning a review into whether that case is true. Sir Jim Mackey's candour has made that contradiction public. What happens between now and the autumn will determine whether ministers correct it before the break clause forces their hand, or whether they are found, once again, defending a procurement decision they cannot fully justify.