Expert sounds alarm bells over upcoming NHS data platform A leading expert has warned that the value of the NHS's Federated Data Platform (FDP) will depend on usability testing if it is to improve patient safety and efficiency in the UK health service. In late December, a month after the FDP was finally awarded to Palantir – the contract for which was heavily redacted – Helen Jones, …
... so obviously the report will be completely ignored.
I think there's also the human element, and over reliance on technology. So this example-
Jones cited a situation where a four-year-old child suffered harm after receiving 10 times the intended dose of an anticoagulant on five separate occasions due to an undetected prescription error recorded on a patient record system.
Could probably have been avoided by someone noticing the dose was rather high. But prescriptions are something of a mess it seems. I'm on a bunch of meds. All of them are 'repeat prescriptions'. But the process is I have to request the repeat, some need to be checked by the GP because they're only allowed something like 3 repeats. So then those meds drop off their prescription request screen and I have to add them to the notes. Then go to the pharmacy, discover those have been missed and request them again. Then because of the delay, the 28 days worth of meds goes out of sync and it's a general PITA.
That seems a simple IT fix. If I'm prescribed repeat stuff, I shouldn't really need to be in the loop other than being informed when my prescription is ready to collect. GP and pharmacist can still check the prescriptions before they're issued.
If I'm prescribed repeat stuff, I shouldn't really need to be in the loop other than being informed when my prescription is ready to collect.
That depends. Some meds you take as needed, so it's not possible to predict when your repeat prescription should be created again and if it is created automatically you may end up with a small stockpile. If the drug can develop an addiction, it may be a dangerous thing.
That depends. Some meds you take as needed, so it's not possible to predict when your repeat prescription should be created again and if it is created automatically you may end up with a small stockpile. If the drug can develop an addiction, it may be a dangerous thing.
That's true, and I think the reason behind the 3 repeat rule. But that's a clinical decision down to the GP and their professional experience. If they have concerns, they can call me in for a review or discuss changing treatment. All mine are supposed to be taken regularly, and include potentially addictive stuff like codeine. And sure, I could decide to skip days to stockpile, but some medications can also be dangerous if withdrawn due to prescriptions being missing.
But as an IT person, it seems a cumbersome and error prone process that's easily fixed. Repeat until loops are, after all one of the most basic software and prescribing tasks.
The report makes good points about Electronic Patient (or Health) Records (EPR / EHR) systems, but this is actually irrelevant to the Palantir contract, because that is for a FDP - a federated data access platform. Every single NHS hospital and GP surgery in the country *already* have EHR systems, but from a number of different providers, so they aren't joined up - the Palantir contract is for a federated data access system enabling easier access across the different systems. The issues with the FDP are concerns about making it too easy to access patient data, not with usability affecting patient care - since clinicians at the point-of-care (i.e. in hospital, GP, or other clinical setting) probably won't be using the FDP directly. Instead it will mostly be used by planners, administrators, and researchers (or that is the idea anyway - obviously there are concerns about it being used for e.g. commercial pharma research, training ML models etc.).
And the core problem is they have overlapping incompatible systems. The NHS Trust system - with individual CEO’s, FD’s IT teams and numerous other duplicated teams is the biggest waste of money and effort out there. It’s worse than the Co-op.
Fake choice, and busy work.
About personal privacy....
One central argument for centralised data is that there are huge benefits for improvements in diagnosis, and then perhaps improvements in treatment.
We are also told that personal privacy is protected because the records are "pseudonymised".
Unfortunately, the Government completely fails to mention some central arguments about risks:
(1) https://www.theguardian.com/technology/2019/jul/23/anonymised-data-never-be-anonymous-enough-study-finds
(2) No mention of "data in transit"..... focus only on a final "central database"
(3) Never mind Palatir SELLING the data, or otherwise abusing trust, what about unknown third parties HACKING the database
Then there's the reliability of each NHS Trust in the first place:
(4) https://www.ft.com/content/6954971e-5d3a-11e9-939a-341f5ada9d40
(5) https://www.theguardian.com/technology/2017/jul/03/google-deepmind-16m-patient-royal-free-deal-data-protection-act
And even Government agencies routinely ignore the law......never mind Palantir:
(6) https://www.theregister.com/2022/01/10/ipco_report_2020/
Yup........It's not just that Palantir is a "contract too far".....Government agencies, NHS Trusts, Palantir................all represent substantial risks.
But El Reg needs to recognise that the basic technology is also a BIG RISK in itself!
Quite a few confused issues in the above post.
1) this is a federated platform not centralised.
2) pseudonymisation is relevant specifically for research or other secondary use of data. For operational or clinical needs which is where the FDP sits the data needs to remain identifiable. So your points (1), (4-6) are not relevant because they all pertain to research.
3) hacking is a risk with any system. Are NHS-managed systems more or less secure than palantir?
Not at all defending the Palantir FDP, but your points aren't where the problems are.
What could go wrong?
I would be interested if Palantir could modify patient data without doctors' knowledge. For instance, let's say there is an undesirable person for the US government living in the UK.
Could Palantir see someone got admitted to hospital, and see they are allergic to certain drug, remove that entry without trace and in effect person dies of allergic reaction?
This entire thing should be scrapped.
Not to worry - they'll be employing a new VP with specialist knowledge of the UK system in 2025, who will be able to assuage your concerns. You'll know he's really good at his job, as they'll be paying him £2M a year despite working only three days a month.
His name?
Mr R. Sunak.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2025
