[NetBehaviour] The NHS and the blockchain

[Edward Picot]

I've now made so many unsuccessful and partially-successful attempts to 
get my head round the Blockchain concept that I'm starting to think I 
might have some form of early dementia.

However, there's one field in which I dimly understand what the 
implications might be: namely, the health service, where I work. Since 
the Labour government introduced the 'Payment by Results' system into 
the NHS about fifteen years ago, and then the Conservative government 
put groups of GPs (clinical commissioning groups, or CCGs) in charge of 
local health budgets, there's been no end of muddle about how to get 
good reliable statistics out of the system as regards which patients are 
being treated where, what they're having done, how much it's costing, 
and which health authorities they belong to. The blockchain is probably 
an answer to this problem.

Let's say a patient rings 999 one weekend because he's having a 
heart-attack. The ambulance takes him to the local A&E department at the 
Tunbridge Wells Hospital. He gets investigated, and after investigation 
he gets transferred to St Thomas's Hospital in London for a triple 
bypass. Then he's discharged to a Cottage Hospital in Hawkhurst, and 
eventually back home.

At each stage of the journey he has incurred costs. First of all the 
ambulance trust charges for transporting him (once to the Tunbridge 
Wells Hospital, then from Tunbridge Wells to the St Thomas' Hospital, 
then back to the Cottage Hospital). Then his A&E attendance and 
investigations in Tunbridge Wells will all incur costs (via the 
Maidstone and Tunbridge Wells Hospital Trust); then his treatment and 
stay in St Thomas's (via the Guys and St Thomas' Hospital Trust); and 
finally his stay in the Cottage Hospital (which comes under the 
Maidstone and Tunbridge Wells Hospital Trust again).

Now, all of these costs and data about what treatments were carried out, 
length of stay, drugs dispensed to the patient while in hospital, etc, 
are supposed to find their way into our local health informatics system, 
which is a big 'data silo', so that if we want to (or, more to the 
point, if the CCG wants to) it's possible to 'drill down', as they call 
it, and find, under the Cardiology costs for a particular financial 
year, the treatment and costs for that particular patient as a result of 
that particular health episode. The difficulty is that the information 
has to be pulled in from various different trusts - our local hospital, 
the hospital in London, and the local ambulance service - and compliance 
varies from trust to trust. So the information from our local hospital 
trust will probably be available more or less straight away, the 
information from the ambulance trust a bit more slowly, and the 
information from London a bit more slowly again. Things can get even 
more complicated if our patient has his heart attack while he's on 
holiday in Dorset - because all the costs he incurs should still come 
back to the area in which he is a registered patient, but of course a 
hospital in Dorset feeds back information much more slowly to Kent than 
it would to its own health authority. And things can also get more 
confusing if parts of the patient journey, while still chargeable to the 
NHS, are carried out by one of the private hospitals - let's say the 
patient, instead of having a heart attack, has a cataract operation at a 
private hospital, which is doing cataract operations on an NHS contract 
as part of the Any Qualified Provider arrangements. The private hospital 
may not have good arrangements in place for feeding back data into the 
NHS system.

A big part of the problem is that you've got all these different 
organisations operating within the NHS - hospital trusts, ambulance 
trusts, CCGs, individual surgeries, private hospitals etc. - and they've 
all got their own bespoke computer systems with their own bespoke ways 
of recording patient data, and it's a constant struggle to get them to 
talk to one another. A blockchain distributed ledger would surely be an 
improvement on the existing system. You'd just have to enter a 
transaction onto the blockchain every time you performed some kind of 
service for a patient - anything from a prescription for paracetamol to 
a hip replacement - and as soon as the transaction was recorded the 
information would be available from one end of the system to the other, 
with the costs correctly allocated both to that particular patient and 
to the patient's own health authority. Of course you'd also have to 
record the same event on the patient's clinical records, in order to 
keep an accurate clinical history, so you'd either have to enter it 
twice, once on the clinical record and once on the blockchain, or (much 
better) you'd have to get every clinical system in the country to 
communicate with the blockchain, which would probably be a lot easier 
than trying to get them all to talk to each other.

So far so good. However, what do you do in the case of a patient where 
you can't discover the NHS number, so you can't accurately say who the 
patient is, where he's registered and where the costs ought to be 
allocated? Let's say somebody's been run over in the street and is taken 
to hospital unconscious, with no identification. Or let's say it's 
somebody from abroad, or a refugee or illegal immigrant who has never 
registered with a GP in this country. One option is to issue a dummy NHS 
number and have some kind of 'miscellaneous' budget against which the 
costs can be allocated. But the other option is to use the system as a 
means of identifying people for whom the NHS doesn't have to accept 
responsibility, and thus excluding or rejecting them. The refugee, the 
illegal immigrant or the person from overseas, who couldn't produce any 
evidence of valid NHS registration, wouldn't be refused emergency 
treatment - not unless there was a really dramatic change of philosophy 
- but if it was anything less than life-threatening they might be turned 
away, or told that they could only have treatment if they paid for it. 
And that's one of the potential effects of the blockchain, as I 
understand it: it's so efficient, that if you set the rules up in a 
certain way at the outset, you'll end up disenfranchising people who are 
misfits of one type or another. If you don't build some leeway into the 
system, you can simply make it impossible for certain types of people to 
get anything out of it. Presumably the same thing could happen to the 
benefits system. And this, in turn, is likely to encourage a black 
market. If you haven't got an NHS number, and therefore you can't get 
treatment, the way round the problem is to steal somebody else's identity.

- Edward