NHS England is to change the methodology behind two of its most widely cited claims about the Palantir-built Federated Data Platform, after internal emails disclosed under a freedom of information request showed the health service's own analysts had questioned the validity of the baseline used to measure the system's success, the Financial Times has reported.
The emails, obtained by the FT under the Freedom of Information Act 2000, concern two headline figures published by NHS England. Firstly, that hospitals using the FDP have delivered 110,000 additional operations, and that they have recorded a 15 per cent reduction in discharge delays. Both metrics are now due to be recalculated using an adjusted baseline method in the next quarterly update, expected in September, 16 months after the figures were first published.
Commenting on a draft of the evaluation methods webpage, one official wrote, "I don't think this is a valid baseline", warning that comparing post-adoption activity with the six months immediately before adoption left the analysis exposed to pronounced seasonal patterns. The official raised the same concern about how discharge delays were calculated for trusts with fewer than 12 months of post-adoption data.
On 11 June 2026, an NHS England official confirmed in an email that the imminent quarterly update did not yet contain the adjusted baseline method for the two metrics, but that the changes would appear in the following publication round. The live methods page was updated on 12 June, with the explanation of the pre-adoption baseline for additional operations no longer present.
NHS England added a caveat to its methods page in June acknowledging that the data did not demonstrate the platform's effectiveness, prompting an investigation by the Office for Statistics Regulation and a further qualification of the claims in late July. The FOI material obtained by the FT shows that caveat followed an internal email of 28 May stressing that the analysis did not establish cause and effect because it was not controlling for all variables.
Charles Tallack, NHS England's former head of operational research and evaluation, told the FT the disclosure raised serious questions about the organisation's analytical culture, and said confidence in its published claims depended on analysis being conducted in line with government guidance.
Ministers face mounting pressure over the 2027 break clause in Palantir's £330m contract to build the platform, which is designed to link disparate NHS data sets. Ben Coleman, a Labour member of the Commons health and social care select committee, described it as shocking that claims about the platform's positive impact were published after staff had raised methodological concerns, and questioned whether an extension of the contract could be justified.
Martin Wrigley, a Liberal Democrat member of the Commons technology committee, said the emails showed the figures could not be relied upon and called for independent verification, arguing Palantir should be replaced in the NHS and its wider role in government examined.
The discharge figures had already come under question after the FT identified errors in a key underlying dataset and the Health Foundation found no noticeable improvement in discharge performance at hospitals using the patient-tracking tool. Separately, Nicola Byrne, the national data guardian for health and adult social care in England, criticised NHS England last week for failing to disclose that a small number of external Palantir staff could access identifiable patient information, warning the omission risked eroding public trust.
NHS England said challenge and continual review were a proper part of its analytical practice, describing the revision as a small change to the baseline calculation for organisations that have used a platform product for less than 12 months.
The FT's report is available here (subscription required): https://www.ft.com/content/bc589991-4e3a-4676-98bb-dcfd0e9e4b9d
