Warner rejects call for CfH architecture review

Tags: CfH Choice Government GP Granger HIS House of Commons iS London N3 Network NPfIT Office Quality Reform Virtual

27 Oct 2006

Health minister Lord Warner has rejected calls from leading computer academics for an independent review of the technical architecture of the NHS national programme for IT.

Speaking at a conference in London yesterday he said: "...I do not support at the call by 23 academics to the House of Commons Health Select Committee to commission a review of NPfITs technical architecture. I want the programme's management and suppliers to concentrate on implementation, and not be diverted by attending to another review."

The 23 academics earlier this month wrote an open letter to the Commons Health Select Committee calling for an independent review of the £12bn NHS IT programme.

In their letter the group urgently called for an independent technical review, describing it as an essential step to help ensure the project succeeds. The group urged the Health Select Committee to carry out “an immediate inquiry to establish the scale of the risks facing the NPfIT".

The 23 leading computer sciences related academics first wrote to the Health Select Committee in April this year expressing their concerns about the technical feasibility and risks associated with the £12bn NHS IT programme, currently running two years behind schedule.

They were subsequently invited to meet with NHS IT director Richard Granger who subsequently invited briefing. CfH and the academics issued a joint statement saying “a constructive and pragmatic independent review of the programme could be valuable”. No such review has since occurred.

Warner said the CfH programme was central to the government’s NHS modernization agenda and had already been vindicated by July’s National Audit Office report. “A positive report was received from the National Audit Office this summer despite subsequent attempts to undermine the objectivity of that report.”

The minister’s speech equated criticism any aspect of the late-running programme as opposition to health service reform and digitization. “Critics of the programme often conveniently overlook the unsatisfactory nature of the current paper-based system.”

In his speech Warner outlined the achievements of the NHS IT programme in delivering infrastructure. Offering a choice selection of key facts about the programme he highlighted the N3 network, claiming it as the “largest virtual private network in Europe”.

He said the planned storage capacity for Picture Archiving and Communications (PACS) – for which 72m images are now held digitally - was 10 petabytes. “Enough to store 1.3m DVD films,” Lord Warner helpfully explained.

He offered Tesco supermarket store openings as a helpful comparator to progress on PACS. “I’m told that Tesco opens one new store week, so we’re actually digitizing hospitals faster than Tesco is opening new stores.”

Warner also said the over 5.6m electronic prescription transfers have now occurred, with 1,023 – about one in ten practices - now using the system. Other achievements highlighted were NHSmail, where over 200,000 users have been registered and the introduction of the GP Quality Management and Analysis System.

Notable by its absence though was any mention of the delays to the systems at the heart of the programme: the national summary and local detailed Care Record Service applications that are meant to deliver detailed integrated electronic medical records for everyone in England.

To date in the secondary care sector the programme, through its prime contractors, has delivered just over a dozen replacement patient administration systems, and a handful of very few clinical systems. Key suppliers have either been sacked or replaced, creating further delays.

Warner acknowledged that not all had gone smoothly: “Given its size and ambition it is not surprising that there are glitches. But overall we are well advanced with delivering the infrastructure of Connecting for Health.”

He, however, restated the government’s commitment for the programme: “"Let me be clear and unequivocal: the Government is committed to ensuring that NPfIT is fully implemented and delivered. We are not going to be deflected by naysayers from any quarter. We recognise that more needs to be done on articulating the benefits that the programme will bring to patients and also to NHS staff."

Readers Comments

Crisis? What crisis?

27 Oct 06 16:05

I hope Lord Warner is right but I fear he is not. There is a history of bad control code at the heart of the systems that have been delivered in the past, by one of the major software suppliers atleast, and that has been known about within the industry for years.

Stories emerging over the summer suggest that this situation still exists. These commercial companies will not release their control code to the universities for peer review, probably claiming commercial confidence but in reality as this would expose these inherent weaknesses to public scrutiny.

Thsese weaknesses will only emerge once the implementations are over and the contract fulfilled. By then it will be too late. NHS organisations complaining that data outputs are not as expected will be told that they haven't configured the software correctly, and that will be that!

Out of touch

27 Oct 06 17:04

"the current paper-based system" ... "digitizing hospitals": phrases which would suggest to the uninformed observer that the health service is lost in an olde world of quills and paper. How flippant to ignore the many excellent and advanced computer systems which were in use before NPfIT was conceived and which, in many cases, continue to be used while CfH stifle criticism and out-of-touch ministers make bizzare comparisons with supermarkets.

Infrastructure acheivements?

mary.hawking@nhs.net

28 Oct 06 07:49

According to your report, Lord Warner mentioned ETP, NHSMail and QMAS as achievements of NPfIT.(No mention of C&B? Wonder why!) I wouldn't know about ETP - it hasn't reached my practice yet although we volunteered early. NHSMail suddenly withdrew the facility to autoforward: in at least one case, this resulted in an Out of Office type answer that this address was no longer being used - please contact on an nhs.uk (insecure) address.. I know a good many people who are *registered* , don't use it - and after not being notified of an impending password expiry, probably can't use it. As for QMAS - the original software was produced by a team at NHSIA before it was abolished: I hope the same team transfered to CfH - but as we haven't got the new updated QMAS yet, rather doubt it!

Does the basic architecture of these supermarkets follow the following fundamentals?

BenSurgison@Hotmail.com

29 Oct 06 11:44

I have done some research into the development of non business specific systems with the aim to produce a scaleable, maintainable, and dynamic system that can change easily both when requirements change mid-development and to take advantage of better technology when it appears. All that needs to be done is to build in the appropriate business logic (PAS systems included!). Note that I have implemented a proof of concept for some of what is described below.

Milestones to a dynamic and scaleable application.

1.Implement both a development methodology and product that can react and respond to change within the market as quickly as possible. Agile processing and SOA (service oriented architecture) is the key.

2.Arrive at a design that is easy to prototype, develop, manage, test and not be tied to a specific platform, environment, scale or even business.

3.The fastest and most cost effective way to produce a product is to base it on an open source framework where reusability, platform independence and scalability are built in.

4.Investigate different open-source frameworks where business-logic can be implemented with ease. Take into account the history of each framework including the current trend of its popularity and implementation successes and disasters.

5.The selected framework must allow the base product to be modularized, not just by separating business logic, but also the different tiers of components such as a data repository, a way of accessing that repository, a way of viewing the data etc.

6.Each module (as described above) within the product needs to be loosely coupled. Nothing should depend on the implementation of each component, or even the location of the component. The only dependency should be the interfaces between each component. This means that as better component products appear in the market-place they can be swapped in and replace the redundant components with no detrimental affect on the system as a whole.

7.Business logic should be able to flow between each component using a standard messaging format such as xml (whether it be hl7, soap, wsdl or pox - plain-old-xml). Any external products should be wrapped within components that provide this interface and therefore hide the implementation of that product.

8.Using this method makes it possible to create a prototype system by using component stubs for each component.

9.The prototype system will not only allow planning and feasibility processes to take place, but also a method of demonstrating the system to any potential customers.

10.The prototype system should also be used as a platform for testing the implementation of each real component as they are implemented.

11.The individual components within the prototype system provide a means of breaking down development responsibility.

12.Each component can be allocated to different teams that can be given responsibility for the development of that component. Each team will have members that can arrange themselves in such a way as to get the task done. They will need to cover design, development and testing of their component and be responsible for the quality of all these steps.

This can be achieved using a combination of the following development methodologies:

XP (Extreme Programming). DSDM (Dynamic Systems Development Method). Scrum.

13.System wide functionality such as security, transaction processing and logging should not complicate business logic. Instead a framework should be implemented that adds this functionality by the use of Aspect oriented programming and cross-cutting concerns. This will allow teams to concentrate on their individual task at hand knowing AOP can be added later.

14.Migration of historical data into a new system by writing directly to the data repository should be avoided at all costs! Instead, the migration should simulate the outer tier of the application entering the data manually in order to make use of validation within the system itself. A system that is based on service oriented architecture is ideally suited for such a task.

15.It should be possible to provide many views of the business logic either via a browser, mobile phone, pda or even inter-process communication such as migration of data.

16.It should be possible to implement a template system to allow the development of such a front end with the collaboration of the customer.

If anyone out there feels that this is a good idea, let me know in your comments!

Even spaceships can make adjustments during flight

29 Oct 06 11:49

Lord Warner feels that things should progress along current lines with CFH, but in all mega-projects there should be room for adjustments when, with the benefit of experience, changes need to be made - but then NHSCFH doesnot appear to listen to the many monitoring groups and people who have been involved over time in informatics to support health. Strident claims that the programme is still 'AOK to go' may result in the CFH ship blowing up. That would be criminally negligent, as there is so much potential if the only the course-corrections were made!

Not a CRS at all.

29 Oct 06 20:04

I have found that if you cannot get it right on paper, you will never get it right on a computer! Getting doctors to agree on a paper system is almost impossible, so forcing them to use a digital system is even closer to impossible, especially if it is difficult to use. Doctors can be made to use computers, when the effort of using the system gives an equal or greater reward in terms of rapidly providing doctors (and nurses etc) with the information that they need to reach diagnoses, make treatment plans, assess progress and if needed, review the diagnoses.

The CRS I have seen makes items like the next of kin's birth hometown and date of birth mandatory (no use in diagnosis at all) whilst leaving fields like "probable" diagnosis (diagnoses) completely non obligatory! So what we are getting is not a CRS, it is a rather clunky, non intuitive, half cooked PAS, which is not as good as the one we have already?

Where is the advantage? Why put in the effort to use it, when currently there is no value for the investment of very expensive clinicians' time to use it - time we should be spending thinking, making professional judgements and very difficult decisions. The computers should aid professional thinking not distract and frustrate.

I suggest an 8 week pause, during which the CRS developers and providrers find out what clinicians and their support staff actually do. Lord Warner seems to think that hernias are repaired with scalpels, not keyhole surgery - he might be just as mistaken about computer systems. I support the call for a review - or the pause!

And then he left ...

03 Nov 06 17:36

It was a shame that Lord Warner skirted all around but did not even mention the detailed Care Record. And then he left the meeting, which was a great shame. The CfH speaker who followed also spouted statistics, carefully avoiding how many NCRS deployments there had been to date.

The Trust Chief Executive Speaker following that discussed the trauma of implementing the IDX Care Reord, with the viability of the Trust threatened by lack of developed reports, and GP referrals dropping until the Choose and Book interface was working. And the prospect of imminently re-deploying to another system.

And speakers from Newham and Homerton then described that 3 years into their Cerner deployment, reporting from the system is still a very difficult process, and the anglicisation and 'hardening' of the Cerner PAS are still underway, only really helped by their direct control over the contract, and being able to flex the system.

Which of course is not available to struggling CfH deployments in the South.

What a joy it would have been had his Lordship bothered to stay and listen to the discussion.

Architecture is not relevant ...

05 Nov 06 12:03

... whether "dynamic and scaleable" or otherwise.

The suppliers are responsible for the development and delivery of the services and can develop them as they see fit.

Has Lord Warner actually tried to use CRS?

06 Nov 06 08:03

I can't think why it has taken me so long to ask this question? Has Lord Warner and the rest of the top people at NPfIT actually tried to use the Cerner Millennium software release zero with order comms that is being "rolled out" into the Southern Cluster? They expect many low skilled minimum wages staff to use it - presumable their IT skills are as good as that? I think that if they were sat down and "trained" and given some bread and butter "PAS" type activities to complete e.g. request a follow up appointment, place a patient on a waiting list for an operation, view a clinic 3 months ahead to cancel and rebook, becasue of holiday, and better still get them to request a set of blood tests along with a urine test. I think that a morning would be enough to convince them that there are real problems!

Architecture is relevant!

06 Nov 06 23:59

If a supplier is trying to shoe horn an offering born from another country that bears no resemblence to the NHS, it would make sense that their architecture was such that it was easy for them to adapt! If however their architecture was such that change is difficult if not almost impossible in a reasonable timescale it would then become apparent when they are either unable to deliver or force the acceptence of something that does not quite fit that which CFH intended. .... Of course that has never happened in this project!