New Minister Comes Out Fighting on Federated Data Platform

Last month, James Frith, the new Parliamentary Under-Secretary of State for Health Innovation wrote a combative reply to Layla Moran setting out the government's position on Federated Data Platform. This was in response to the letter Moran wrote on behalf of the Health Select Committee who had conducted a hit job on FDP in June, inviting Martin Wrigley MP onto the panel to execute his campaign against Palantir, resulting in a letter recommending the break clause be exercised. Anyone who reads Frith's letter will see it for what it is: a clear and complete rebuttal of the committee's recommendation to exercise the break clause. However, the Guardian reported Frith's letter as evidence that Palantir mistrust is driving patients to opt out. Digital Health reported it as proof the NHS is "in a great place to ditch Palantir." Neither outlet reported what the letter actually says, and it follows that we can't trust these media to tell the accurate story. So what does the letter say?

Manchester is not an alternative.

Frith tells the committee that the Greater Manchester system "is an analytics platform and cannot enact specific actions related to direct patient care (for example recommend patients for discharge, an operation, or providing a national vaccination service)." It operates within one ICB. Scaling it nationally is "unproven." This was written three weeks before Martin Wrigley told the Prime Minister at PMQs that Manchester "supplies more than Palantir does or ever can." The minister's letter contradicts Wrigley's preposterous claim. Frith adds that while Greater Manchester ICB refused to adopt FDP, every NHS Trust in Greater Manchester has adopted it and is using it operationally.

The benefits methodology was defended, not withdrawn.

The committee said the evidence of benefits is contested. Frith responded that the benefits methodology "is aligned to the green book methodology for major programmes and has been highlighted by independent assurance teams as an example of excellence." The Office for Statistics Regulation investigated and "did not find significant defects in its evidence which would lead NHS England to withdraw the evidence it has published on the value of the NHS FDP." An independent evaluation is running at Imperial College London. The NAO is also reviewing. This is a minister telling the committee that challenged the benefits that the methodology is sound, the regulator found no grounds for withdrawal, and independent evaluation is under way. That did not make the headlines either.

The NHS owns everything.

All products, data models, analytical solutions, and intermediate artefacts are NHS intellectual property. Palantir is forbidden from using them commercially. The CDM and operational products would persist after the contract ends. If the NHS owns everything, what would a replacement supplier actually replace? The answer is Foundry, the software layer underneath. It would be like replacing Excel and keeping your own spreadsheets.

A replacement by March is unlikely.

Frith confirms that normal government procurement timelines make it unlikely a new contract could be in place by March. The exit clause that Foxglove and Digital Health celebrated requires Palantir to provide up to 24 months of transition support. That means Palantir keeps running FDP for two more years while the NHS simultaneously procures and builds a replacement. The NHS pays for both. That is not ditching Palantir. That is keeping Palantir while spending additional money on something that does not yet exist, that would most likely not be delivered within that timeframe.

The opt-outs track to the campaign, not the platform.

Frith writes that there was a small spike of around 20,000 when FDP was announced in 2023, and that "it has only been with increased media interest that we have seen an increase in opt-outs over the past four months (~60k)." He describes this as modest. The underlying data tells a different story: the baseline rate was roughly 400 to 600 opt-outs per week for nearly two years, accelerating to 7,500 per week once Good Law Project began encouraging opt outs on social media. And the national data opt-out does not apply to FDP. It prevents data being used for cancer research, drug development, and clinical trials. Sixty thousand people withdrew from medical research because they were told to be afraid of a platform the opt-out does not touch.

Frith assembled all of these facts in a single letter. Manchester cannot replace FDP. The benefits methodology is sound. The NHS is not locked in. No replacement is ready. The opt-outs are driven by campaigns on social media, not by the platform. Every outlet that covered this letter chose to report something other than what it says.

The letter is public. I would encourage anyone with an interest in this debate to read it and draw their own conclusions.