This week ministers decided to extend Palantir's controversial contract for the NHS Federated Data Platform (FDP) by two years while the Government procures a UK replacement. Looking for a British supplier is the right call. Health data is one of the country's most valuable assets, with sovereignty over health data becoming a bigger concern both in the UK and overseas.
But sovereignty is only one of the problems with the current platform. The FDP's whole-life cost has risen to over £1bn, yet the NHS owns none of the software or intellectual property. That leaves the NHS dependent on a single overseas supplier. Frontline reviews are also mixed; more than 20 trusts have stopped using the FDP's waiting list tool, with some saying the information doesn't link back to their patient administration or electronic patient record systems.
Simply replacing the supplier, however, is not enough. A UK company running the same model where patient data is copied into a supplier's cloud would leave the NHS with the same security and lock-in risks. Here are four things the new tender should look for.
1. Never move patient records
Today, data from each trust is copied into Palantir's cloud to be processed. Each trust decides who can access its data, but the data itself is centralised. The centralised data now includes test results, imaging and histology reports, and GP records from a five-practice pilot. Every copy of these sensitive records is another point of exposure to cyber-attack or misuse. Copies also further entrench the supplier, because the more data one company holds, the harder the company is to replace.
To avoid another billion-pound lock-in, the new tender should not allow patient data to be held nationally at all. Only operational data, such as bed capacity, theatre schedules and staff rosters, should be pooled, and it should sit on cloud infrastructure the NHS owns and controls rather than the supplier. The supplier's software would run inside that NHS environment, on the NHS's terms. Patient records should stay with the trust or GP practice that holds them. With the right technology, frontline workers can still have the experience of a single platform, with a patient's appointments, notes and test results in one place, while the underlying records never move to an external supplier.
If patient records stay with trusts and GP practices, and operational data sits on NHS-owned cloud, the NHS keeps its data and environments even if it ends the contract. Switching provider would mean changing the software layer on top, not starting another major infrastructure project. Organisations such as Genomics England already use this model: its data stays in its own secure environment and the supplier's software runs inside it.
2. Judge every app against an NHS target
The criteria for the 2023 tender were heavily shaped by Palantir's earlier work and the results since the platform launched have been mixed. Almost a million people have been removed from waiting lists, yet another source reports 13 of the 41 trusts using the FDP's theatre scheduling tool reported fewer operations after adopting it.
Every tool built on the new platform, from theatre scheduling to bed management, should be tied to a named NHS target. Each should be tested against a baseline, with results published trust by trust. Tools that don't meet their target should be improved or withdrawn.
3. Earn consent from trusts and patients
More than 44,000 people have lodged legal objections to their data being used on the FDP. Complainants are using Article 21 of the UK GDPR, the right to object, meaning NHS England must stop processing their data unless it can show legitimate grounds to continue.
Patients shouldn't have to lodge a legal objection to be heard on how their medical records are used. The new tender should require a clear plan for earning public trust from the start. Patients should be able to see who uses their data and why, for example in the NHS App, and choose how it is used.
NHS trusts deserve the same choice. They should join because the platform helps them, not because they are pushed. Clinicians should also be involved in designing the apps, so they fit the working day rather than adding another system to it.
4. Build on what trusts already run
Many trusts and ICBs have invested in building their own data platforms, and some are already well adopted. Greater Manchester Trust told MPs its own technology can match the FDP's functionality. Leeds Teaching Hospitals warned that switching to the FDP risked a loss of functionality for users rather than a gaining.
The new tender shouldn't waste that investment. Any new supplier must work with the systems trusts already use, through common NHS rules for how systems connect and share data. That way, trusts keep the tools they've built, staff don't have to learn a new system, and if the NHS ever needs to change supplier, it won't be starting from scratch.
Sovereignty is moving up the agenda
The extension to Palantir's contract gives the NHS two years to rethink what it needs from its data platform. That's long enough for a handful of trusts to test different approaches, and for the NHS to judge the results against its targets before the next contract is signed. British technology is already proving huge benefits for national health programmes overseas. It's time the NHS benefited.
