NHS England has commissioned an independent review into the effectiveness of Optica, the patient discharge tool built on technology supplied by the American company Palantir. The move follows analysis from the Health Foundation, a think tank, which found no meaningful difference in discharge delays at trusts using the system. That conclusion directly contradicts the figures NHS England has cited for more than a year to justify the wider rollout of Palantir's software across English hospitals.

Optica forms part of the Federated Data Platform, a £330 million programme awarded to Palantir in 2023 without competitive tender. The platform is intended to pull together fragmented NHS datasets, tracking bed capacity, patient flow and discharge coordination in a single system. Optica itself was developed under a separate series of contracts worth around £60 million before being folded into the wider platform. Its stated purpose is to help hospital staff identify patients ready to leave and clear the administrative obstacles that keep beds occupied longer than necessary.

Bed-blocking, the term used for patients who remain in hospital after they are medically fit to be discharged, has remained one of the most persistent pressures on the NHS. Delayed discharges tie up beds needed for new admissions, contribute to lengthy accident and emergency waits and cost the service money that ministers are keen to recover. Software promising to ease that bottleneck has therefore attracted considerable political attention. NHS England has previously said Palantir's technology helped deliver 110,000 additional operations, a 15 per cent reduction in discharge delays and a 6.8 per cent rise in patients learning within 28 days whether they need cancer treatment. Those figures were repeated by ministers and officials and appeared on the health service's own website.

The Health Foundation's analysis challenged that record. Its researchers examined discharge data across trusts using the platform and found no notable improvement once other factors were taken into account. Separately, an investigation by the Financial Times identified errors in the underlying hospital data behind the 15 per cent figure, with four trusts confirming mistakes after reporters spotted sudden and unexplained drops in recorded discharge delays that later reversed. Irregularities were found across 42 per cent of NHS hospital trusts over four years of acute discharge data. Charles Tallack, NHS England's former head of operational research and evaluation, said the evidence for the platform's impact on discharge delays looked increasingly flimsy, noting that the dataset used may be adequate for day-to-day management but was not built for formal evaluation.

Faced with this, NHS England has moved from public confidence to a more cautious position. The Office for Statistics Regulation had already told the health service to add stronger caveats to its published figures, making clear that before-and-after comparisons do not establish causality. NHS England has now gone further, commissioning Imperial College London to evaluate the Federated Data Platform, with particular attention to Optica, where officials acknowledge the question of causality is most difficult to resolve. NHS England has said the results of that evaluation will be published.

The episode sits within a broader pattern of scrutiny facing large technology contracts in the public sector, where claims of efficiency savings are often made before independent verification is possible. Palantir's expanding footprint in Britain, which now extends to defence and policing work alongside its NHS contracts, has added a political dimension to the debate. Some MPs and campaign groups have called for the £330 million contract to be reconsidered altogether, citing the company's links to the American and Israeli defence sectors alongside the disputed performance data.

The Imperial College review will not settle the political argument over Palantir's role in the NHS, but its findings will carry weight. A conclusion that Optica delivers little measurable benefit would raise awkward questions about how the health service assesses new technology before committing public money, and about how such claims are verified once contracts are already in place.

‍