Abandoned Nhs It System Has Cost £10Bn So Far

[interestrateripoff]

http://www.theguardian.com/society/2013/sep/18/nhs-records-system-10bn

An abandoned NHS patient record system has so far cost the taxpayer nearly £10bn, with the final bill for what would have been the world's largest civilian computer system likely to be several hundreds of millions of pounds higher, according a highly critical report from parliament's public spending watchdog.

MPs on the public accounts committee said final costs are expected to increase beyond the existing £9.8bn because new regional IT systems for the NHS, introduced to replace the National Programme for IT, are also being poorly managed and are riven with their own contractual wrangles.

When the original plan was abandoned the total bill was expected to be £6.4bn.

Excellent news for GDP. All this money being spent and nothing being delivered. With all this money we could give "free" school meals to all kids at least their might be something to show for it at the end.

I wonder how many people are on 6 figure salaries not delivering anything with this!

[scrappycocco]

10 billion lol, I'd like to know who got what in the back pocket.

[NorthamptonBear]

10 billion lol, I'd like to know who got what in the back pocket.

Follow the money. Every time it will show how and why things happen.

[frank rizzo]

It would be nice if they did have some kind of patient records system than the ancient system they currently have.

Last year I had an operation on an old ankle injury. After a referral to physio from my GP I spent the first visit re-explaining why I was there and the history of the injury to the therapist who then sent me away with an exercise sheet and an appointment for the next week.

This went on for about six weeks before they agreed to send me for a scan and appointment with a specialist. At that appointment (many months later) I re-explained for the third time what my problem was as once again the person I went to see had no prior knowledge of my complaint.

At the next meeting my paper records were there, including notes from a hospital visit nearly 20 years ago! Full details of my problem were available to all these people all the time and none of them ever bothered to check.

I ended up waiting nearly 18 months for an operation I knew I needed from the start and they would have too if they'd bothered to check the notes.

[fluffy666]

Step 1: Scan everyone's records.

Step 2: Shove the results through OCR. You'd want the scanned files as well for when it went wrong

Step 3: Create a cloud-based folder for everyone. Add the documents.

Step 4: Add a simple search facility (optional)

[swissy_fit]

Step 1: Scan everyone's records.

Step 2: Shove the results through OCR. You'd want the scanned files as well for when it went wrong

Step 3: Create a cloud-based folder for everyone. Add the documents.

Step 4: Add a simple search facility (optional)

The NHS drops it's pants, lubes up and twerks vigorously in the direction of one or two IT services companies for all these projects. I wonder who the shareholders are?

[Bloo Loo]

snip

I ended up waiting nearly 18 months for an operation I knew I needed from the start and they would have too if they'd bothered to check the notes.

the best system in the World doesnt get people to actually read the notes...my recent experience proves this...at Spinal triage, the physio was interested in the note I filled out in the waiting room...I pointed out that I wasnt sure why I needed to fill in all this stuff as the letter from the Physio referring explained it all...even then she was about to cut my need for MRI until I repeated and emphasized I had a serious fall and was in constant pain directly following the fall..

"Ah, you had a trauma that resulted in the symptoms?."

YES....

"I see, well lets get you booked in then"

You see, the form I filled in didnt even ask how I thought I got the issue in the first place. She had the referall, copius notes and the new form Id just filled in....

[Bemusedmover]

the best system in the World doesnt get people to actually read the notes...my recent experience proves this...at Spinal triage, the physio was interested in the note I filled out in the waiting room...I pointed out that I wasnt sure why I needed to fill in all this stuff as the letter from the Physio referring explained it all...even then she was about to cut my need for MRI until I repeated and emphasized I had a serious fall and was in constant pain directly following the fall..

"Ah, you had a trauma that resulted in the symptoms?."

YES....

"I see, well lets get you booked in then"

You see, the form I filled in didnt even ask how I thought I got the issue in the first place. She had the referall, copius notes and the new form Id just filled in....

You describe the problem perfectly....................

a file contains the complete history of the patient............

the health care "professional" only reads page 1

...............of course we assume that the "professional" actually "has" a qualification

.............Oh and can actually be bothered to read!

............even the BBC has noted how few qualified staff are actually on duty at any given time

I suggest you ALL hope that you are NEVER ill!!!!!!!!!!

[ChumpusRex]

Step 1: Scan everyone's records.

Step 2: Shove the results through OCR. You'd want the scanned files as well for when it went wrong

Step 3: Create a cloud-based folder for everyone. Add the documents.

Step 4: Add a simple search facility (optional)

It's a bit more difficult than that, in reality.

1. Need to make sure that records are correctly filed in relation to a particular patient. Even if they are initially misfiled, or a new record is created in error, the patient changes their name, an administrator incorrectly combines a pair of case notes, but then needs to undo their mistake, etc.

2. Need to be in a standard format, so that they can be transferred to another site.

3. Must be able to interact with, query and cross reference data with other information systems (e.g. national demographics database, departmental results servers, etc.).

3b. Must be able to automatically update records, when an authoritative system updates key data (e.g. if a patient's name is changed on the national ID database, that will need to feed forward to the records system)

4. Needs to provide some sort of structure to the file - medical progress, nursing progress, laboratory results, reports and formal letters.

5. Needs to restrict access to authorized staff, provide an audit trail of access, with the option for multiple levels of confidentiality.

6. Must be capable of preserving the full dataset provided, e.g. robust industry standard colour management for photographic images, etc.

There are other issues, simply scanning paper notes really doesn't work well. You get a free-document system but without the ease of use of a properly filed paper file with section markers, etc. If you want to make it electronic, it really needs to enforce a sort of structure.

Electronic medical records are a mature product line, having been available worldwide for about 15 years, with many software packages on their 6th or 7th release. They all suck badly in one way or another, because the vendors have not fully appreciated the degree of complexity in delivering a product which is robust enough, flexible enough and supporting enough communications formats to actually work.

[Bemusedmover]

It's a bit more difficult than that, in reality.

1. Need to make sure that records are correctly filed in relation to a particular patient. Even if they are initially misfiled, or a new record is created in error, the patient changes their name, an administrator incorrectly combines a pair of case notes, but then needs to undo their mistake, etc.

2. Need to be in a standard format, so that they can be transferred to another site.

3. Must be able to interact with, query and cross reference data with other information systems (e.g. national demographics database, departmental results servers, etc.).

3b. Must be able to automatically update records, when an authoritative system updates key data (e.g. if a patient's name is changed on the national ID database, that will need to feed forward to the records system)

4. Needs to provide some sort of structure to the file - medical progress, nursing progress, laboratory results, reports and formal letters.

5. Needs to restrict access to authorized staff, provide an audit trail of access, with the option for multiple levels of confidentiality.

6. Must be capable of preserving the full dataset provided, e.g. robust industry standard colour management for photographic images, etc.

There are other issues, simply scanning paper notes really doesn't work well. You get a free-document system but without the ease of use of a properly filed paper file with section markers, etc. If you want to make it electronic, it really needs to enforce a sort of structure.

Electronic medical records are a mature product line, having been available worldwide for about 15 years, with many software packages on their 6th or 7th release. They all suck badly in one way or another, because the vendors have not fully appreciated the degree of complexity in delivering a product which is robust enough, flexible enough and supporting enough communications formats to actually work.

The biggest problem is that this is a database...............

upper and lower case entries completely change the parameters..............

a simple problem that IT continues to ignore :angry:

you IT guys are really ...............well what are you?????

[ChumpusRex]

The biggest problem is that this is a database...............

upper and lower case entries completely change the parameters..............

I've not found that in practice.

The one that is really common is the apostrophe problem. If the text you have typed contains an apostrophe, the data cannot be saved, and you get some random "database command error". I've not seen this in any 2013 editions software packages, but it was a big problem in some 2010 editions.

On a similar vein, "special characters" continue to cause big problems even in the latest H2 2013 software. The ampersand character often cannot be transmitted between systems, so for example if I were to write "The patient was taken immediately to A&E for assessment", the message would be broadcast to other record systems in a garbled and illegible format.

[Oliver Sutton]

Mark Easton tweet.

Since 2009 in England, NHS hospital doctors pay has risen 5.5%, nurses pay 7.5% & senior managers pay 12.9%

Reward for failure.

[Guest eight]

It's a bit more difficult than that, in reality.

1. Need to make sure that records are correctly filed in relation to a particular patient. Even if they are initially misfiled, or a new record is created in error, the patient changes their name, an administrator incorrectly combines a pair of case notes, but then needs to undo their mistake, etc.

2. Need to be in a standard format, so that they can be transferred to another site.

3. Must be able to interact with, query and cross reference data with other information systems (e.g. national demographics database, departmental results servers, etc.).

3b. Must be able to automatically update records, when an authoritative system updates key data (e.g. if a patient's name is changed on the national ID database, that will need to feed forward to the records system)

4. Needs to provide some sort of structure to the file - medical progress, nursing progress, laboratory results, reports and formal letters.

5. Needs to restrict access to authorized staff, provide an audit trail of access, with the option for multiple levels of confidentiality.

6. Must be capable of preserving the full dataset provided, e.g. robust industry standard colour management for photographic images, etc.

There are other issues, simply scanning paper notes really doesn't work well. You get a free-document system but without the ease of use of a properly filed paper file with section markers, etc. If you want to make it electronic, it really needs to enforce a sort of structure.

Electronic medical records are a mature product line, having been available worldwide for about 15 years, with many software packages on their 6th or 7th release. They all suck badly in one way or another, because the vendors have not fully appreciated the degree of complexity in delivering a product which is robust enough, flexible enough and supporting enough communications formats to actually work.

Maybe we should all just look after our own?

[tomandlu]

The biggest problem is that this is a database...............

upper and lower case entries completely change the parameters..............

a simple problem that IT continues to ignore :angry:

you IT guys are really ...............well what are you?????

Err... no. Case-dependency is an option (and we turn it off on ours which has both positive and negative effects - some coding systems, for instance, differentiate codes on the basis of case).

[tomandlu]

We provide specialist renal database applications, and the truth is that no generic system will handle the speciality-specific data that we handle. In addition, each system needs to be bespoke to the individual hospital, since there is no agreed standard of workflows between units or what detail of treatment needs to be captured beyond a few basics.

Anyone working with NHS IT Systems had a very dim view of this whole business. A laudable aim that was completely beyond the scope of current technology.

[thecrashingisles]

The market capitalisation of CSC, one of the main contractors, is about $8bn. It would have been cheaper for the government to buy them out!

[Mr Jib Fingers]

It's a bit more difficult than that, in reality.

1. Need to make sure that records are correctly filed in relation to a particular patient. Even if they are initially misfiled, or a new record is created in error, the patient changes their name, an administrator incorrectly combines a pair of case notes, but then needs to undo their mistake, etc.

2. Need to be in a standard format, so that they can be transferred to another site.

3. Must be able to interact with, query and cross reference data with other information systems (e.g. national demographics database, departmental results servers, etc.).

3b. Must be able to automatically update records, when an authoritative system updates key data (e.g. if a patient's name is changed on the national ID database, that will need to feed forward to the records system)

4. Needs to provide some sort of structure to the file - medical progress, nursing progress, laboratory results, reports and formal letters.

5. Needs to restrict access to authorized staff, provide an audit trail of access, with the option for multiple levels of confidentiality.

6. Must be capable of preserving the full dataset provided, e.g. robust industry standard colour management for photographic images, etc.

There are other issues, simply scanning paper notes really doesn't work well. You get a free-document system but without the ease of use of a properly filed paper file with section markers, etc. If you want to make it electronic, it really needs to enforce a sort of structure.

Electronic medical records are a mature product line, having been available worldwide for about 15 years, with many software packages on their 6th or 7th release. They all suck badly in one way or another, because the vendors have not fully appreciated the degree of complexity in delivering a product which is robust enough, flexible enough and supporting enough communications formats to actually work.

Not exactly beyond the wit of man though is it?

In computer entertainment they are doing things far far more complicated with much smaller teams. For example Eve Online which was built by a small team is a massive multiplayer game that has complex physics, in game economy, thousands of stats updating in real-time and tens of thousands of players interacting simultaneously. Bit more complex than having some data stored and updated discretely with a decent user interface. How exactly they spent £10bn on that is anyones guess.

[Tin Foil Hat]

Let me assure you all it has not cost anything like this as they are only paid on delivery and a lot of it never made the grade. A lot of companies went bust as they couldn't deliver.

[tomandlu]

Not exactly beyond the wit of man though is it?

In computer entertainment they are doing things far far more complicated with much smaller teams. For example Eve Online which was built by a small team is a massive multiplayer game that has complex physics, in game economy, thousands of stats updating in real-time and tens of thousands of players interacting simultaneously. Bit more complex than having some data stored and updated discretely with a decent user interface. How exactly they spent £10bn on that is anyones guess.

This comes up time and time again - games are full of bugs and security issues that wouldn't be acceptable in an NHS records system. In addition, Eve may be complex, but it defines its own limits and rules; it doesn't have to capture reality.

[Dorkins]

1. Need to make sure that records are correctly filed in relation to a particular patient. Even if they are initially misfiled, or a new record is created in error, the patient changes their name, an administrator incorrectly combines a pair of case notes, but then needs to undo their mistake, etc.

5. Needs to restrict access to authorized staff, provide an audit trail of access, with the option for multiple levels of confidentiality.

True

2. Need to be in a standard format, so that they can be transferred to another site.

3. Must be able to interact with, query and cross reference data with other information systems (e.g. national demographics database, departmental results servers, etc.).

3b. Must be able to automatically update records, when an authoritative system updates key data (e.g. if a patient's name is changed on the national ID database, that will need to feed forward to the records system)

4. Needs to provide some sort of structure to the file - medical progress, nursing progress, laboratory results, reports and formal letters.

6. Must be capable of preserving the full dataset provided, e.g. robust industry standard colour management for photographic images, etc.

All of this sounds nice, but not actually necessary for the task of getting information about the patient's history to the doctor who is treating them. Scanned records with the correct NHS number attached would give you 80% of the benefit for 20% of the cost and would at least be a good starting point to build from in the future.

[Mr Jib Fingers]

This comes up time and time again - games are full of bugs and security issues that wouldn't be acceptable in an NHS records system.

The game data has to be secure otherwise hackers would ruin the game within days. If people could hack the game to give themselves extra stuff or affect other players then there would be no game.

Yes there are small bugs in animations and so forth, but you don't really need that sort of thing in an NHS database.

In addition, Eve may be complex, but it defines its own limits and rules; it doesn't have to capture reality.

The NHS database just needs to store some patient information that you can recall doesn't it? You set the limits of what data you want to store and categorise it accordingly. I would hardly say that is capturing reality.

[tomandlu]

You set the limits of what data you want to store and categorise it accordingly.

No, reality sets the limits. If something's inconvenient to code for in a game, you just don't code for it. You don't have that option when you're trying to encode real life.

[Mr Jib Fingers]

No, reality sets the limits. If something's inconvenient to code for in a game, you just don't code for it. You don't have that option when you're trying to encode real life.

So you're saying the data isn't categorised before being entered in these databases at all? Or it is and that increases the complexity by a factor of about a thousand to account for the extra cost?

I'm not saying there aren't problems for these types of projects, just there are large comparable problems in games as well such as networking, frame rate, graphical effects, physics systems, 3D rendering, multiple platforms and graphics cards, texture memory, localisation etc. etc. I just can't see how having difficult data to manage is going to cost 1000s of times as much to solve as these problems.

[motch]

So you're saying the data isn't categorised before being entered in these databases at all? Or it is and that increases the complexity by a factor of about a thousand to account for the extra cost?

I'm not saying there aren't problems for these types of projects, just there are large comparable problems in games as well such as networking, frame rate, graphical effects, physics systems, 3D rendering, multiple platforms and graphics cards, texture memory, localisation etc. etc. I just can't see how having difficult data to manage is going to cost 1000s of times as much to solve as these problems.

they should of got the GTA 5 team in..

[fluffy666]

It's a bit more difficult than that, in reality.

1. Need to make sure that records are correctly filed in relation to a particular patient. Even if they are initially misfiled, or a new record is created in error, the patient changes their name, an administrator incorrectly combines a pair of case notes, but then needs to undo their mistake, etc.

2. Need to be in a standard format, so that they can be transferred to another site.

3. Must be able to interact with, query and cross reference data with other information systems (e.g. national demographics database, departmental results servers, etc.).

3b. Must be able to automatically update records, when an authoritative system updates key data (e.g. if a patient's name is changed on the national ID database, that will need to feed forward to the records system)

4. Needs to provide some sort of structure to the file - medical progress, nursing progress, laboratory results, reports and formal letters.

5. Needs to restrict access to authorized staff, provide an audit trail of access, with the option for multiple levels of confidentiality.

6. Must be capable of preserving the full dataset provided, e.g. robust industry standard colour management for photographic images, etc.

There are other issues, simply scanning paper notes really doesn't work well. You get a free-document system but without the ease of use of a properly filed paper file with section markers, etc. If you want to make it electronic, it really needs to enforce a sort of structure.

Electronic medical records are a mature product line, having been available worldwide for about 15 years, with many software packages on their 6th or 7th release. They all suck badly in one way or another, because the vendors have not fully appreciated the degree of complexity in delivering a product which is robust enough, flexible enough and supporting enough communications formats to actually work.

The thing is, I wasn't talking about a database.. that's a hard problem.

Having them scanned and online, based on personal accounts, in the way I said is the first step and has the huge advantage of being doable.

Indeed, you'd need to do it just to get a starting point to start even thinking about creating a database.