« Faint Praise | Main | Factoid of the Day »
May 31, 2006
Computers in the Health Service
We’e known for some time that the NHS all singing, all dancing computer system is way over budget. Now a Minister has finally admitted it.
A computer system for storing the medical records of all 50 million health service patients will be more than two years late and three times over budget at £20 billion.
While that number is spread across some years, it is in fact equal to some 1.5 to 2% of one year’s GDP. Yes, really, a system for medical records and appointment setting is costing some 2% of the entire country’s output for one year.
There are others out there who will give far better detailed descriptions of what’s gone wrong than this technophobe can but just from the newspaper report, a few ideas:
Lord Warner also said the real cost of the NHS programme for IT (NPfIT) was likely to more than three times the previously stated £6.2 billion total - something admitted by officials 18 months ago but not so far owned up to by ministers.
There’s the admission.
The scheme is the biggest civil information technology programme in the world.
That is at least part of the problem. Just as there are economies of scale so are there diseconomies.
Responding to the criticism expected in next month's NAO report that those running the scheme had failed to deal with the concerns of NHS staff, he said: "That is probably a fair criticism in part.
"We possibly could have got into the game earlier, and we could probably have done it better earlier on.
Some years into the programme and only now are they taking an interest?
"The National Care Record plan involves electronic summary medical records - possibly including major diagnoses, operations, recent test results, current medications and allergies - and more detailed records to which access will be limited to regional medical staff." A pilot study is due to begin next year.
Some years into the programme and they are still using words like "possibly"? "Pilot"?
The British Medical Association's family doctors committee has said patients must be asked for their consent before their records are computerised.
Rather than such an "informed consent" approach, the Government's current preferred option is "assumed consent" under which people will be assumed to be happy to have their details stored electronically unless they specifically opt out.
A spokesman for Connecting for Health, the government agency in charge of the NPfIT, said: "We note the position of the BMA GP committee but the medical profession is not united on the issue. We've adopted the opt-out model."
They still haven’t decided the most basic rules of the structure? No wonder the thing is late! C’mon guys, it shouldn’t be this difficult. No one starts building a bridge by throwing up a few bricks on hte side of the river and then wondering about how to link them all. You plan first, everything is designed in minute detail, even the length of each and every piece of steel, before you even start.
And just think of the implications for the cost of the ID Card system, which will be even larger. And more complex. 5% of GDP anyone?
May 31, 2006 in Web/Tech | Permalink
Comments
"everything is designed in minute detail, even the length of each and every piece of steel, before you even start"
Not in a big IT project it isn't. Because by the time you get to implementation, the cost ratios have all changed. And some (most) of the tools you aimed to use are obsolete. Data design concepts, however, haven't changed much in 30 years. That should be nailed down.
Posted by: dave heasman | May 31, 2006 9:42:21 AM
But ID cards will cost £6bn, not £19bn like the LSE said. Oh no. The ID card project will run to budget, because it's being run by.. the Home Office.
K.
Posted by: Kay Tie | May 31, 2006 10:38:27 AM
Some doctors have (rightly IMO) expressed concerns about the threat to patient confidentiality but I don't think that is sufficiently widely appreciated. All sorts will be working on the development and implementation of this massive IT system and it's unclear how widely access rights will be distributed. While I don't think there is anything on my personal medical records that could cause me embarrassment, that is not true for everyone. STDs? An abortion? Potentially, targeted data mining could extract rich seams of information for blackmailers.
Any public IT project billed as "the largest" in the world is a good candidate for public concern. How have other countries managed their healthcare systems without it? The last I saw, the NHS had a fairly modest place in World Health Organisation rankings for effectiveness - above America certainly, but below the placing of a string of other west European countries. Over the last 50 years to my knowledge, the NHS has often been politically sold to the electorate on the basis that the American system is awful but few comparisons are made between the NHS and healthcare systems in other west European countries. Curious that.
Posted by: Bob B | May 31, 2006 10:42:33 AM
I have a radical proposal. Every Doctor's surgery can book appointments and keep patient details using a little known piece of software called microsoft outlook. If they include a useful piece of information such as the patient's NHS number then anyone in the NHS can quickly obtain details on a patient by sending something called an email to the Doctor's surgery. I know that MS Office is pretty untested but it might be worth the risk.
Posted by: Mark T | May 31, 2006 11:21:02 AM
Ignoring whether this is desirable or not and the consent issues, surely this is a simple network and database/data warehousing project.
There are several of nationwide networks with multiple points of access and different levels of user input and access. They also have some very high levels of security. Without them the entire domestic banking industry would grind to a halt.
As for the data, there are any number of off-the-shelf database and data warehousing products that can be panel beaten to suit almost every situation known to man. The amount and type of data to be stored will simply dictate the size and number of file servers needed.
If even a dumb-ass mining engineer knows enough about IT to understand this, our "lords and masters" must be truly ignorant beyond the point of incompetence.
And I would suggest that Plod (or at leasst some accountants and IT experts apointed by Plod) takes a look at the tenders and quotes provided by whichever company won the contract.
RM
Posted by: The Remittance Man | May 31, 2006 11:37:50 AM
RM,
Quite right. I mean, £20 billion? What the hell are they doing...?
DK
Posted by: Devil's Kitchen | May 31, 2006 11:56:22 AM
Lets say there are 60 million patient records.
With a decent index (LogN(6*10^8)) it comes in at about 18 comparisons to find a patient record. I could run the whole thing on my home P.C. (slightly higher spec than average ahem).
It's the insane OVER-speccing (in order to extract medical stats) that is killing this. Just bung everything into one text and let the doctor read it.
Can I have my 2 Billion now?
Posted by: AntiCitizenOne | May 31, 2006 1:22:30 PM
Well, there's more than 60 million = 6 10^7, since I doubt your record is deleted when you die, and searches go like log to base 2 so 26 or thereabouts, but it is broadly correct to say that the data volume is of the order of magnitude of any number of civil database projects. Google's data warehousing requirements must be far, far bigger than this. The figures that have been bandied about are simply beyond belief. Are we to assume that this is the development cost alone, or does it include implementation and deployment? Even so a billion pounds will buy you a lot of computers and infrastructure (in fact it will buy you several dozen supercomputers). That there is so little public outrage at the squandering of such a grotesque amount of money, and that no heads have rolled, is hard to fathom.
All software projects that fail do so because of shoddy management. This is a problem in the private sphere to be sure, but bureaucrats and politicians are intrinsically incapable of handling IT management as the incentive structure for successful delivery of a system is completely different from that which obtains in the commercial world. This is another example of government messing with things it is inherently ill-equipped to provide.
Posted by: David Gillies | May 31, 2006 5:03:45 PM
"This is a problem in the private sphere to be sure, but bureaucrats and politicians are intrinsically incapable of handling IT management as the incentive structure for successful delivery of a system is completely different from that which obtains in the commercial world."
That diagnosis was evident at least several years back - not least because Computer Weekly, the leading industry periodical on the markets and jobs, was keeping score and publicising failing and failed government IT projects. As best I can tell from afar, there have been attempts at remedies, mainly by shifting more of the burden of technical and implementation risks on to the developers:
http://www.bjhc.co.uk/news/1/2006/n605005.htm
However, I suspect some politicians incline to meglomania. I am not reassured when I hear on the BBC what amount to boasts that the NHS IT project - Connecting for Health - is the "biggest IT project in the world", as though that is something we ought to wonder at and admire when what really counts is successful outcomes for patients.
The problem of meglomaniacal visions tends to be endemic in the NHS. I recall from ten years ago seeing official NHS literature which was proclaiming (correctly) that the NHS is the largest single employer in the whole of western Europe. The interesting issue is why did the NHS hierarchy consider it worthwhile making that claim in official PR/information literature distributed to the public? Are most people impressed and reassured by the claim?
Posted by: Bob B | May 31, 2006 8:34:14 PM
I've worked on some massive IT systems with some very complex processing.
I know roughly how much they cost to build, and this dwarves them. If you think about the cost ratio between a Kia Picanto and a Ferrari 575, you still wouldn't be close.
Posted by: Tim Almond | May 31, 2006 11:12:00 PM
The government was forewarned clearly enough four years ago by a report in The Economist (2 May 2002) detailing failed and failing government IT projects and pointing to the downstream implications for a mooted NHS IT project: "Your health depends on it"
"IF GORDON BROWN is still basking in the apparently widespread approval for his plans to revive the National Health Service with a massive transfusion of cash, a small item of seemingly unrelated news last week, reported in Computer Weekly, should worry him. A leaked memo from the Lord Chancellor's department indicated that Libra, a £319m IT project designed to link magistrates' courts with other organisations, such as the police, customs and the Crown Prosecution Service, was on the brink of collapse. Libra's problems are the latest in a long list of government IT horror stories (see table) which bode ill for the government's ambitions. NHS reforms, like all public service improvements, depend heavily on IT. . . There is now a raft of Whitehall-wide guidelines aimed at trying to learn from past mistakes. These include: the need for projects to be 'owned' by one responsible and competent individual; recognition of the fact that projects are often over-ambitious and need to be broken down into bite-sized chunks; the realisation that the way people work has to change when new technology is installed; the importance of drawing up contracts that have the right mix of sticks and carrots to make vendors deliver on their promises; and, above all, the need for clear and precisely-defined objectives. . . "
http://www.economist.com/displaystory.cfm?story_id=E1_TTTTJJD
Posted by: Bob B | May 31, 2006 11:19:24 PM
Having worked as a developer for a tiny small-scale government project, i can say that the money is probably all going to development so that the stakeholders can change their mind, then change their mind again, then change it again and again, until they've got some FrankenProject nothing like what it should have been. The scope creep in government is appaling...
Posted by: anon | Jun 22, 2006 11:28:25 AM
Argh, politicians always gives people too much false claims and promises. I doubt that they'll know how does a computer or database really works.
Posted by: Young Engineer | Sep 4, 2006 7:25:24 AM
