The critique of the US data model
In a letter published by the Guardian, Professor Caroline Mitchell, a general practitioner in Sheffield and a research lecturer in general practice at Keele University, weighs in on the case of the alleged benefits of Palantir's artificial intelligence tools for the NHS. According to the signatory, the news that NHS England may have overstated the efficiency gains linked to Palantir would not surprise many frontline clinicians. The reference is to an article dated 12 November, which reported that NHS England had withdrawn a statement described as misleading about the number of trusts that would benefit from the US company's artificial intelligence tools.
The central argument of the letter is that global digital policy too often tends to assume that software designed for the US healthcare system—fragmented and insurance-based—can simply be adapted to a universal public service. Mitchell calls it a legitimate concern that the algorithms underlying decisions made with artificial intelligence may be immature in the NHS context, having been developed mainly for a very different healthcare system.
The purpose of the systems: billing versus patient care
In support of her position, the author cites two accounts. A consultant physician at a local trust told her that he had to work across seven additional software systems for a single consultation, one of which was dedicated to billing functions. A senior US hospital executive, by contrast, explained to her that the main purpose of most healthcare information systems in the United States is to record care for reimbursement purposes. The same executive said he was astonished that, by contrast, NHS general practice records are designed to support patient care throughout life and safe prescribing, and that the NHS app is oriented toward health coding rather than using up an annual insurance cap.
The letter also raises a question of technological sovereignty: why would the NHS need a US company to manage its data when the United Kingdom has a strong tradition of innovation in healthcare software? Mitchell cites the case of Emis, developed under the guidance of general practitioners, which transformed primary care computing decades ago. The proposed conclusion is that software built for an insurance market may simply not be the best fit for universal health systems.
What remains to be verified
The material does not contain Palantir's or NHS England's position in response to the letter, and does not indicate whether the news about the withdrawal of the statement on the 137 trusts has been the subject of official clarification. Technical details are also unavailable on the Palantir systems used by the NHS, the contracts in place, or the results of efficiency assessments: elements necessary to establish whether the issues raised by Professor Mitchell have been encountered in practice. The letter should therefore be read as a qualified opinion piece, based on reported clinical and administrative experiences, and not as an independent verification of the tools' performance.
