Is the Federated Data Platform too expensive to cancel?
- 19 September 2023
Sticking with a hugely expensive plan to develop a ‘mythical beast’ of a fighter plane in the 1960s did not end well for the government of the day. The Federated Data Platform may be similarly doomed, writes Joe McDonald
Two of my kids are junior doctors. I often tease them that the two hardest selection processes in the world are medical school entry and fighter pilot training, and I passed in both. An attack of political insight in my teens steered me away from the place I was offered in the RAF fast jets training programme in favour of medicine. But make no mistake, I remain an Aerospace anorak.
As a boy in the 1960s, the next big thing was going to be the BAC TSR-2. Experts were split at the time on whether manned fighter planes were already made obsolete by high-speed heat seeking missiles or whether we just needed better faster planes. A hugely ambitious spec was written for a plane that would replace both the Spitfire and the Lancaster Bomber. It would operate at high altitude and low altitude, high speed, low speed, drop nuclear bombs and do ground attack, all while being six metres longer than the Lancaster and considerably heavier.
The issue of whether this plane for all seasons was pie in the sky or the great leap forward became politically polarised. As the budget for this mythical beast spiralled out of control and test flights were underwhelming, the Labour opposition used it as evidence of a profligate Tory waste of money that post-war Britain could not afford. For the Tories, it would set Britain apart and make the British Aerospace Corporation the world’s best aircraft manufacturer – if they pulled it off.
They spent so much money on it that it became too expensive to cancel. The Tories lost the election and the new Labour Prime Minister, Harold Wilson, immediately cancelled TSR2. Every penny spent was wasted.
Cut from the same cloth
Now there is another project on the runway with a price tag of nearly half a billion pounds. The Federated Data Platform (FDP) has twice been the subject of the Digital Health Debate Series, most recently on Friday last week, where a very knowledgeable audience moved from 60:40 in favour of the programme to 51:49 against it. Quite a swing. Why?
During the debate it became obvious to that the FDP and TSR2 were cut from the same cloth. The FDP, like the fighter plane, would solve all the problems we could think of, but nobody could explain exactly how or when, or even what the problem was that it was supposed to be solving. It wasn’t clear what we are getting for our half a billion pounds over seven, yes seven, years.
There was no clarity about whether it was for direct care or research. There was no clarity about how citizens would be consulted about their health record being shared with bookies’ favourite in the procurement race, the controversial, Trump-supporting, and openly hostile to the NHS, AI peddlers, Palantir. There was no clarity about the data model to overcome – the unacknowledged GIGO elephant in the room.
There was some acknowledgement from those in favour of FDP that the comms has been suboptimal, but I’m sorry that doesn’t wash. This is not the first time NHS England has gone for a massive national data grab and the comms is always absent because they know what will happen if they tell people what’s going on. Emboldened by Covid we are clearly in the middle of Care.Data 3.0.
Normally this goes the same way. The project’s sponsors will press ahead, a great deal of public money will be wasted and the project will ultimately fail, dying on the altar of people who are passionate about privacy but who were not properly engaged and who function as canaries in the privacy mines for the general public. The general public will take their opinion from those privacy experts and not from the government.
National vanity project
However well-intentioned the FDP undoubtedly is, some of us have put 10 years of work into gaining citizen’s trust to deliver regional shared care records so that the IG weather has changed from ‘no sharing’ to ‘sharing by default’. We have moved slowly and with citizens’ consent. The Local Shared Care Records community are within an understandable ‘diameter of trust’ for the citizen, the government is not. We have respected privacy campaigners and not dismissed them as ‘tin foil hat wearers’. The worry with this latest national vanity project is that it is now politically too expensive to cancel, just like TSR2.
I predict the contract will be awarded in the next few weeks, and it will be awarded to Palantir. In the ensuing furore two things will happen. Firstly, a great opportunity to improve the basics of NHS IT will disappear in a half billion pound puff of AI hype. My junior doctor children will continue to have to “fight six nurses for a turn on the ward PC”.
Secondly, there will be a further loss of trust in the NHS. Every time the government decides they are trusted with the citizens’ personal health record, people opt out of information sharing, which might affect their health. Actual patient harm in return for a set of benefits from FDP which have not been convincing to a very well-informed crowd in the Digital Health Debate. What will Joe Public make of it? What will this seven-year project look like after the next general election which is 18 months away. What might a new Prime Minister make of it? Will there be a pause for thought?
Professor Joe McDonald is medical director at Sleepstation, SARD JV, Parsek and Ethical Healthcare Consulting
10 Comments
“Trump-supporting, and openly hostile to the NHS, AI peddlers, Palantir. ”
Karp is more or less opposite Thiel. More than likely PLTR is the best solution if you actually want next-gen system (of course fair to argue one wants to go with simpler IT improvements)
Like do you even know Palantir? Sure they started in US gov / intelligence and somehow that is very bad. (Yes, granted, would be kinda better if EU company… maybe) But they are *so* *so* *so* much more. I would implore you to start searching twitter for examples of Foundry and AIP.
(And yes, small time Danish investor here, so maybe I am just blind bla bla bla)
“Like do you even know Palantir?” To be honest , Thomas I don’t. But as a tax payer I expect my government to explain them to me in return for half a billion pounds of our money.
The one in Dave Kelsall’s cpmment
I think Dave meant this: https://www.theverge.com/2018/2/27/17054740/palantir-predictive-policing-tool-new-orleans-nopd
Thanks! Not sure what went wrong there!
Thanks Dave, broken link.
Which link is broken?
I’m with Joe on this.
When FDP started there was clear water between FDP and shared care records – ShCRs were for direct care, FDP for analytics. Any bridge between the two had to have a clear legal basis with CAG – the Confidentiality Advisory Group – guarding that bridge.
Somehow over time that’s changed and the FDP comms narrative now states “the main use of the Federated Data Platform will be for individual care, so the data for everyone who receives care in England will go into the platform, regardless of whether you have a national data opt-out in place”.
So details of your abortion, your mental health crisis, whatever will go into the FDP database with the patient not having any say, except if they ask the professional caring for them that it not be shared, placing the onus – and the burden – on them to make that call. Of course there will be controls on who can then access it, but if even the PSNI and the Met Police can’t keep their data secure…..
We simply do not learn from our past. Trust takes years to build but can be broken in a moment. The exemplary work of OneLondon public engagement programme showed that well informed patients understand the value of sharing for both direct care and analytics and are generally content but only if they can control whether to share. They may never exercise that control, but the fact that it’s there builds trust.
And although I absolutely get the potential value of data, at a time when we have crumbling RAAC buildings, a workforce taking industrial action over pay and some antiquated basic IT, is it really the best use of scare resources when we already have invested in tools which can do the job ?
Joe recalls TRS2; I’m left here thinking HS2, with a sprinkling of PPE.
FDP is more than just a data privacy issue. FDP is core to the implementation of the Integrated Care Boards. You may have been shown how to suck eggs at some pointlesssly expensive management consultants powerpoint presentation. ICB’s are central to the further privatisation of the NHS under the Health and Social Care Act 2022. ICB’s are a secretive disaster happening as you held your debate. https://calderdaleandkirklees999callforthenhs.wordpress.com/2023/09/16/we-do-not-like-these-ics-we-think-theyll-kill-our-nhs/comment-page-1/#comment-7899
Joe,
That’s a great article, the only thing I’d argue with is “However well-intentioned the FDP undoubtedly is”.
Why would anyone assume that the FDP is well-intentioned – at least when all the dice seem loaded in favour of one potential bidder?
From the perspective of so many people outside the UK Central Government, Palantir is a data-mining solution from a company with dodgy links to elements within the USA’s espionage and law enforcement communities.
Let’s not forget that this is the same solution which has prompted this sort of article in the US
https://d1wqtxts1xzle7.cloudfront.net/56430372/Palantir_has_secretly_been_using_New_Orleans_to_test_its_predictive_policing_technology_-_The_Verge-libre.pdf?1524772881=&response-content-disposition=inline%3B+filename%3DPalantir_has_secretly_been_using_predict.pdf&Expires=1695153487&Signature=FRmEvnEdt50bd0B3EreBIGB7PununcVtXAJtm6RL9CbxX9dcD6sVkWA0pQiM5yNnP7josjapWWF8O9dLUliKZTwZod7J-kaO9rvjLyT~Lbl2VWg1A549PB-3Q2zoTrivuPxoaepUPx5SuYOPQDfG4ZAGv6lPGVhasgbcOisPy1yfbV0LKQQNRg0clk1NNADDLc13~NesWpIQVgxzB0FHRD7iOo~CyzomfWXvprFSAHPbO54fp0EPOdPifQqk7Ix~VRa8KxsFNegr-9PjFgNSpWu6oAhz23npLau1T-6qhXDMU3VdbccHdD6XDuea3RkF4~LCnuol71jz2VhoBnERTg__&Key-Pair-Id=APKAJLOHF5GGSLRBV4ZA
Best wishes
Dave
Comments are closed.