Thread by @marcus_baw

As you all know I'm not one for being overly critical of drivel. But I'm going to make an exception and challenge every inaccuracy in this (well-meaning) article. (thread)

theguardian.com/commentisfree/…

With all due apology to Martha Gill of the Observer who wrote it. I do think there is some good intention here, but the problem is that: Healthcare is complex. Technology is complex. Healthcare technology is very complex. Trying to simplify it leads to utter bollocks.

"A unified database could be a medical gamechanger"

Well, that's very debatable in itself, but let me spell this out:

The NHS Federated Data Platform will not be a unified data platform.

Does the word FEDERATED not give that away?

If you are a health / tech journalist and you aren't aware of the differences between

* direct care
* research
* planning/logistics
* population health

Then my advice is: don't write Observer articles about them, because you will mix them up worse than primary school paint pots

Firstly I challenge the assertion that "you can't do things without [Big Tech]"

Of course an org like the NHS can do this, it has a budget of £billions. It's just failed to even try.

Also: there are options like @OpenSAFELY which aren't "Big Tech" that can do this - properly

Things the NHSE #FederatedDataPlatform will not address AT ALL #1

Everyone understands the need for more joined up care in the NHS. The FDP is being sold to the gullible as a solution to this. It is not. It isn't designed to be that. It hasn't been procured for that.

Things the NHSE #FederatedDataPlatform will not address AT ALL #2

The FDP will not help "the right file" 🤣🤣🤣🤣 from you GP get to paramedics at the point of care. This is Direct Care and is expressly NOT what the FDP is for.

Things the NHSE #FederatedDataPlatform will not address AT ALL #3

The NHS FDP will not help "your vaccine record" follow you as you move cities. This is already dealt with by a system called GP2GP and although that system is old, it works.

Give me ONE example of a "rich country" which has completely solved health data flow.

Most of the biggest health economies are WAY behind the UK in terms of data fragmentation. Add in a layer of billing (as most countries have) and data flows get really hard.

Things the NHSE #FederatedDataPlatform will not address AT ALL #4

Research already happens, and has done for decades. That The Observer was not previously aware of it doesn't mean it wasn't happening.

Again see @OpenSAFELY for privacy-preserving ways to do Big Research

Billions *have* been lost, yes

They *do* always run up against the same problem, yes, which is: Incompetence and lack of actual healthcare technology knowledge in our senior NHS IT leadership.

Good example:

Last year Simon Bolton addressed a room full of Chief Clinical Information Officers and Health CIOs last year (room has about 10 THOUSAND combined years of NHS IT Experience) with this opener:

"I didn't realise Health Tech would be so hard"

Well, there are thousands of NHS IT professionals who DO understand health tech, DO know it's hard, and have been DOING it for years - rather than being tapped on the shoulder.

Why not teleport one of them into a senior NHS IT leadership role?

OK so on to the opt out.

Here the lack of an opt out is presented, unchallenged, as a way to make better research.

Is PATIENT CONSENT not a consideration here? Not even in the bloody Observer?

Perhaps the best bit of the whole article. A copy-and-paste synopsis of Palantir as a company. But at least it's vaguely accurate

Peter Thiel: "Rip up the NHS and start over"

The @ICOnews doesn't have the power to "sue" anything "into oblivion". It has the power to apply a fine which an organisation like @PalantirTech can simply regard as the cost of doing business, should it decide to transgress.

The Cambridge Analytica scandal is instructive.

Things the NHSE #FederatedDataPlatform will not address AT ALL #5

Nope, once again, repeating the trope that the FDP will "join up" any of the direct care in the NHS is just dangerous lies. It is repeating NHS England's FDP propaganda. And it is not true.

I don't agree that an organisation of the size and budget of the NHS "was always going to need help from the private sector" - that is a Tory-era neoliberal self-fulfilling prophecy peddled by the unimaginative and clueless.

In 15 years of NHS IT work I've never heard of this person or this company. What credentials do they have to start making pronouncements about what the NHS is or isn't capable of?

There is no doubt that the NHS COULD run data centres if it had a technically capable leadership.

So, if a stupid project is about to commence, and we can see it is stupid, the argument is that we should let it happen because of vague promises of vague benefits in the future?

No lost years, no lost treatments, no lost lives. That is utter claptrap. The FDP is not FOR that.

Congratulations. This sentence is correct. The only one so far.

The concern is that the UK Govt have been playing fast and loose with medical data for years and this FDP gives them even better means to do it.

Delivery (or the failure of) is not actually going to be Palantir's problem. It'll be the NHS's, just like it was when the National Project for IT failed and the suppliers SUED THE NHS AND WON.

digitalhealth.net/2018/09/depart…

In Government IT, delivery is completely optional.

Palantir won't deliver any of the purported benefits - not even the ones the project is meant to deliver.

Much less can it deliver "joined up NHS data", and perfect direct care delivery. Because it was never intended for that.

In the closing paragraph, and well after resigning, Sinom Bolton does actually get it right.

It's only a shame he was never in a position to influence this, as the mere CEO of NHS Digital during much of this period.

I once again apologise to the author of the article for this tirade of abuse. And of course, should any follow-up article be in the offing, I'd be more than happy to spend the half-hour or so on the phone necessary to get the fact straight. Or - talk to @medConfidential