How we revealed NHS England claims about Palantir’s FDP delivering benefit for the NHS are dubious, cherry-picked and should come with a serious health warning
By Foxglove Head of Strategy Dr Tim Squirrell.
In March, a number started doing the rounds: 100,000. This was the claim being made by Palantir, and NHS England, for the number of extra operations being delivered due to the introduction of the Federated Data Platform (FDP). 100,000 is a good number – weighty, but not unbelievable; ideal for helping convince people that Palantir – a firm better known for its work with the US military and ICE – is delivering for the NHS.
But that number – a central plank of Palantir’s marketing campaign as they face the possibility of being kicked out of the NHS – concealed far more than it let on.
The number was presented in media interviews and on the NHS England website with little evidence to support it. We weren’t told how different NHS trusts were contributing to this figure. We were denied any comparison with how hospitals were faring that hadn’t started using the relevant tool. So I stuck in some Freedom of Information requests and waited.
The results were startling: in one column, the names of each trust (41 in all) that had adopted the ‘Inpatient Care Coordination Solution module’ of the FDP, which is meant to optimise how operating theatres are used to increase the number of procedures delivered. In the other column, the number of additional operations. Except, instead of the plus sign to indicate how many additional operations had beencreated, some of those numbers had a minus in front of them. Some 13 of the trusts – about a third – had delivered fewer operations since they started using the tool.
To try to present Palantir’s FDP as a success story, the Government used a methodology that was already dubious, as it excluded the possibility that any other factors were behind the results. They took the average number of weekly operations for a hospital in the 6 months prior to deployment, then multiplied that by the number of weeks the tool had been used to get the number of operations they would ‘expect’ to have happened since the tool went live. Finally, they subtracted that from the number of operations that were actually delivered, giving a number for the ‘increase’ in operations.
The trouble is, even this approach didn’t provide good news all round: 30% of trusts have delivered 9,073 fewer operations since they started using the FDP. The Government’s solution was to aggregate those results into invisibility, cherry-picking their superficially positive figure of 100,000+.
We can’t say with confidence that the FDP caused hospitals to deliver fewer operations. That would be unfair. We don’t have any comparison data for how hospitals that aren’t using the FDP performed, and we don’t know all the other variables that might have affected how many ops could be delivered.
But this is exactly the approach that the Government and Palantir are taking, albeit in reverse: we’re being asked to believe that anything good that happens in these NHS trusts is thanks to Palantir. Anything bad is because of something else.
The risks posed by Palantir are clear: this Trump-supporting US tech giant is focused, in its own words, on global dominance. The BMA have warned it is not a suitable partner for the NHS; a cross-party committee of MPs has warned the Government against being “at the mercy” of Palantir.
We’re told by the firm, and their supporters inside and outside Government, that we shouldn’t worry about this, because the benefits to the health service are so great. But the reality is that unlike the clear risks, the benefits are murky at best – and, as we’ve shown, their claims simply don’t hold water.
NHS England and Palantir cannot, and should not, say that the data they have made available proves FDP caused hospitals to deliver more operations.
Instead, the number of times I’ve had to hear Palantir’s UK head Louis Mosley repeat “110,000 extra operations” (it’s increased since March) without any context is frankly far too many.
Palantir has been very keen to have the argument that they are all about delivering for the British taxpayer while their opponents are ideologically driven zealots. They seem much less keen to interrogate whether they really are delivering.