Part of the Christine Connelly’s job is to defend the NPfIT. Her boss is Sir David Nicholson, the overall Senior Responsible Owner for the NPfIT.
Before Nicholson was appointed Chief Executive of the NHS, and Connelly appointed the first CIO for Health, it was made clear that their support for the NPfIT was part of the requirement.
When, at a press conference at the HQ of the Department of Health last October, I asked Connelly whether she would defend the £5.8bn spent thus far on the NPfIT as value for money, her reply left no room for doubt.
“The money we have spent so far I believe has been value for money,” she said. Some elements are excellent, she added.
“We want to make the best of what we have got and leverage it If you go across the world and talk to other people in other countries they are very envious of some of the things we have in place. If you look at digital imaging - PACS - it is a great system. It works very well. If you look at N3 [national broadband network] which we would talk about as something quite simple to do, actually 10 years ago that didn’t look simple to do at all. We take it for granted but it is a great development.”
But what of the parts of NPfIT that are not going well?
Richard Bacon MP a long-established member of the Public Accounts Committee, wrote to Connelly on 13 January 2011 asking for her assurance that renegotiations of contracts with BT and CSC under the NPfIT could be termed value for money.
In his letter to Connelly Bacon said:
“I am writing following your appearance at the Public Accounts Committee and our subsequent exchange of correspondence because of my continued concerns about the renegotiations of contracts with BT and CSC in relation to the National Programme for IT.
“While I welcome the Government’s decision to seek additional savings from the National Programme for IT, and its commitment to permit Trusts more autonomy in making future IT decisions, I am very concerned about the evidence now emerging about the details of the contract negotiations, and specifically about whether proposed spending reductions are being achieved by permitting BT and CSC to make disproportionate cuts in planned deployments and in the functionality of systems on offer.
“I note your observation in September 2010 that “the best way to deliver value is to honour the contracts”. It would seem to me therefore that we should explore whether value is in fact being delivered for the NHS and for taxpayers. I would therefore be grateful if you could confirm or comment on the following observations ”
Bacon asked Connelly several times whether the facts, as he outlined them, represented, in Connelly’s opinion, value for money.
From Bacon’s letter:
“We should explore whether value is in fact being delivered for the NHS and for taxpayers ”
“ is this [BT’s new contract] really delivering value?”
“ can you say whether it [Cerner Millennium] is still value for money?”“How can this difference in price [between the cost of RiO as supplied by BT and RiO supplied outside the NPfIT]?”
“Does this [proposed memorandum of understanding with CSC) represent value for money?”
Connelly’s replies gave none of the specific assurances Bacon was seeking on value for money.
Yet Connelly’s letter to Bacon was 2,300 words. Clearly her letter was written in good faith, and she wished to be helpful. Others may say that Connelly’s letter shows how civil servants can answer an MP’s questions at length, without answering the main point.
So we still don’t know whether the Department of Health has signed a re-negotiated £546m deal with BT which represents poor value for money but which can otherwise be justified at length.
Indeed Bacon remains concerned that, of the £546m payment to BT, £400m cannot be justified.
Connelly's letter to Bacon, in full.
This was Christine Connelly’s letter to Bacon, with Bacon’s questions in italics:
Director General of Health Informatics and
Chief Information Officer for Health
79 Whitehall, London SW1A 2NS
Mr Richard Bacon
MP for South Norfolk
Member of the Public Accounts Committee
24 January 2011
Re Contract negotiation in the National Programme for IT in the NHS
Dear Mr Bacon
Thank you for your email of the 13th January 2011. This letter seeks to respond to the points you have raised in the order you raised them.
However, I would firstly like to set some context for the actions the Department is currently taking and the Government’s plans for the longer-term future for NHS IT.
Ministers are committed to creating an environment which encourages a vibrant supply market that provides choice of high quality systems that meet common and demanding standards, supporting delivery of high quality services for NHS organisations and the patients that they serve.
Self-evidently, we do not start that process from a blank sheet and have to operate within the constraints which exist because of the contractual arrangements made under the previous administration, the existing level of maturity of the supply market in health IT, and the reality that the structure of health provision in England will continue to change over time.
As Secretary of State, Andrew Lansley has looked at the options available and concluded that we should continue substantially to honour our contractual commitments made under the National Programme for IT, but should seek to do so on the basis of modified delivery arrangements to increase flexibility, and a more modular approach, to enable benefits to be delivered sooner. This approach has been reviewed and supported by the Cabinet Office Major Project Review Group.
Thus the transition to the Government’s longer-term vision for delivery of services will be delivered through a series of milestones over a period of time. Decisions made in the meantime will need to take account of the existing environment, as well as future aspirations.”
Turning now to your questions:
[Bacon’s questions in italics] In London, the new contract signed with BT in April cuts £112m from the original £1.1 billion contract signed in 2003. Of the 32 acute trusts in London, around half will not now get Cerner Millennium as BT will offer at most 10 more Cerner deployments to add to the 8 already in place; fewer community trusts will get RiO; and the requirement for BT to deliver around 1,500 new GP systems across the capital has been dropped. New software for the London Ambulance Service has also been abandoned. This would seem to amount to delivering half the original contract for a small reduction in charges. In your opinion, is this really delivering value?
Connelly: You are right to say that the number of acute Trusts scheduled to receive the Cerner product has been reduced,” replied Connelly. “However this quantitative change needs to be considered in the context of the qualitative, ie functional, enhancements and improvements that have also been agreed. Previously, the Cerner Millennium solution was more limited and the delivery approach being taken had led to a number of concerns from the NHS (and indeed the supplier).
The organisation of the NHS has changed and is anticipated to change further and even more radically in the future. Thus the NHS need was for a highly configurable, more modular approach to delivery of the product, supported for longer in the deployment and rollout phase.
This included the need for additional supplier manpower to support the deployments as well as significant new technical environments and infrastructure to allow the local configurations to be supported.
This locally driven approach seeks to support the new health agenda and was supported by the NHS London advisory groups who reviewed and contributed to the revision of the contract to meet their stated needs.
In respect of RiO, the revised contract, which necessarily accommodated the London PCT re-organisation resulted in one less deployment of RiO for Community Health Trusts from that originally contracted.
It should also be noted, as with Cerner, the revised deal resulted in a number of improvements and enhancements to the RiO solution to be provided by BT.
You are correct the GP systems will no longer be provided by BT, and these will now be delivered by the GP Systems of Choice [GPSoC] programme. The BT contract did scope out a London Ambulance Solution and this could still be “called down” if required and if funding were made available.
However it was determined by the NHS that there are currently higher priorities and the Ambulance funding was therefore directed towards these.
Bacon: I note that Imperial College, contracted to implement Cerner Millennium, recently announced that a rival product has gone live across the entire Trust. Will the NHS pay for the Cerner Millennium deployment at Imperial even if it is not taken up? In the circumstances, can you say with any confidence that BT’s other commitments in London will in fact be taken up?
Connelly: You ask about the alternate solution being deployed at Imperial [Imperial College had contracted to implement Cerner Millennium but later announced that a rival product had gone live across the entire Trust].
Imperial College Healthcare NHS Trust has indeed recently rolled out a legacy PAS [[patient administration system] across the whole organisation, but this is a result of the merger of constituent parts of the Trust, namely St Mary’s acute Trust and Hammersmith acute Trust. These now operate on a single solution rather than two different solutions as they had when entirely separate organisations.
The next planned phase, under the BT LSP contract, is to roll out the Cerner Millennium Order Communications system on top of the PAS system. This will provide benefits to the Trust. The single PAS now deployed makes for an easier Cerner Millennium deployment in the future as St Mary’s and the Hammersmith have already cleansed their historical data being migrated to a single, rather from two, different systems. Subsequently Imperial will rollout the Cerner PAS under the BT contract to replace the single legacy PAS referred to previously.
This approach was anticipated as part of the revised BT contract (the more flexible approach to delivery mentioned earlier) and there is currently no expectation that Imperial will not take up the full planned Millennium deployment.
I am currently entirely confident that BT’s other commitments will be taken up in London, subject to BT continuing to deliver an acceptable service. NHS London have confirmed at the time of the changes to the contract that there was at least the historic level of demand for the services, and I am unaware of any change in that position.”
Bacon: As the functionality of Cerner Millennium has been cut back drastically and, as its name implies, the product is now out of date technically, can you say whether you consider it is still value for money? I would point out that, under the original plan, the products now being installed should have been running live for six years.
Connelly: I am afraid you are misinformed about the level of Cerner functionality. Functionality has not been reduced. Indeed the exact opposite is the case, and at the request of NHS in London, access to the full scope of the Cerner product has been made available for those Trusts where it has or will be deployed to meet their clinical and patient needs. There is currently an agreed programme of activity to upgrade all acute Trusts onto the latest Cerner Millennium codebase. In fact, all live Trusts in the South will have been upgraded to the latest Cerner Millennium codebase by the end of January 2011. The first of the London acute Trusts is scheduled to complete the code upgrade project by the end of March 2011, and the rollout will continue across London throughout 2011.
Bacon: In his letter to me of 18th November 2010, the Comptroller and Auditor General stated that BT will be paid £36m for each of the seven NHS trusts running Cerner Millennium already installed by the Southern region’s former local service provider, Fujitsu.
It will get £23m for each of three new Cerner installations. And it will receive £9m for each of the 25 RiO community and mental health sites in the South.
All prices are said to be based on pricing in the other LSP contracts. However, it has been reported that RIO has been implemented in other Trusts for between £0.5 and £1 million. Is this correct? If so, what questions does this raise in relation to value for money?
Is it also correct that BT will be paid more for delivering Cerner to three ‘greenfield’ sites than it would have been paid for the four sites originally agreed? The Cerner prices seem to be much higher than the prices paid by Trusts such as the Wirral who purchased directly. How can this difference in price be justified?
Connelly: It is difficult to comment meaningfully on the financial comparisons you make about the RiO product, since it is not clear what scope, duration of terms apply to Trusts that you say have taken delivery of RiO for between £0.5m and £1m and how, if at all, these compare with the product for which we have contracted. The price paid for RiO in the South was based on pricing in the BT LSP contract, which has been validated through the Department of Health’s usual business case approvals processes at the time of change of subcontractor.
In respect of the ‘greenfield sites’ in the South of England, I am aware that you raised a number of queries with the National Audit Office and they are providing a full analysis based on advice and information provided by the Department. I therefore do no feel that it would be appropriate for me to comment on this particular issue until the work with the NAO is completed.
As in the case of RiO, so for locally-commissioned Cerner systems. It is not clear to me what scope, responsibilities, duration or terms apply to the contract in the Wirral but in the South BT provided an updated detailed proposal on the basis outlined above, which was priced at £69m. Similarly to RiO, and as you would expect, this figure was agreed through the usual business case processes, involving the Treasury, and our own and Treasury Ministers. This is also currently being reviewed by the NAO, I understand at your request.
Bacon: In the case of CSC and its software contractor iSoft, it would also appear that the companies may secure guarantees of substantial payments even if in practice NHS Trusts choose alternative products, so that there are many fewer deployments than planned.
Connelly: As I believe is widely known, as part of the original contracts, volume commitments were made to CSC (and all other LSPs) to ensure best value for money was obtained. This is balanced by the contractual commitment that payment for product is only made when the product has been deployed and is working.
Thus if the supplier fails to deliver, the planned payments are reduced. If a Trust takes an alternate solution which could have been delivered by the incumbent LSP, then in those circumstances it is appropriate that the supplier is recompensed. This has not occurred in my time as Director General for Health IT.
Bacon: I understand that the DH is negotiating a Memorandum of Understanding with CSC that would see 187 NHS trusts assumed to take the Lorenzo system; only 35 fewer than specified in the original £3 billion contract.
But would you not agree that there is good reason to think that many fewer Trusts will in practice agree to take the system?
In March, Trusts across the North, Midlands and East of England were advised that the version of Lorenzo offered for deployment would have greatly reduced functionality. A number of Trusts have already indicated that they are not committed to Lorenzo and reserve the right to make other arrangements.
I note that in October 2010, you stated that: ‘those who don’t [take Lorenzo] can do what they want and they won’t get penalised, however if we do not get enough trusts in totality to meet the contract then the NHS as a whole will be penalised.’
Does this represent good value for money?
Connelly: The Memorandum of Understanding you mention has not been agreed (and therefore not signed at this stage), and I therefore cannot comment on your speculation as to the precise numbers. The volumes we have been discussing with CSC have been based on the advice from the local NHS via the SHAs last year. As I explained in my earlier response, before any revised contract were signed, we would review and confirm any volume commitments with the local NHS.
Bacon: As you know, the Lorenzo system has repeatedly missed announced deadlines since the 2005 iSoft Annual Report, which stated that it was “available from early 2004".
Most recently, in April 2009, you set a deadline of March 2010 for Lorenzo, stating: "If we don't see significant progress ... then we will move to a new plan for delivering infomatics in healthcare". Yet again, this deadline was missed.
Connelly: I have made no secret of the fact that I regard the history of delivery of Lorenzo to date as very unsatisfactory. Because of this, we have been looking to achieve a more flexible approach to the delivery of the system on the basis that the latest plan from CSC for the delivery of the product, which was accepted last year, is achieved, as well as securing the reduction in contract cost. The Chief Executive of the NHS and the President of CSC are planning to meet to discuss this matter further. It would be inappropriate for me to make more general comment at this stage that might be detrimental to the Department’s negotiating position.
Bacon: Even in the small number of cases where the current version of Lorenzo (much reduced from its original planned functionality) has been deployed, for example at University Hospitals of Morecambe Bay NHS Trust, there have been serious problems.
The Trust failed to go-live with Lorenzo in March 2010, finally went live in the summer, and has been subsequently working on a ‘stabilisation plan’. By December 2010, there were believed to be hundreds of remaining bugs in the system.
I also understand that NHS Bury has yet to sign off the deployment verification for the implementation of Lorenzo Release 1.9 - a year after going live with the system.
Connelly: “We continue to work with CSC and Morecambe Bay, along with other Trusts where it [Lorenzo] has been deployed, to address any defects in the Lorenzo system. Where a deployment has not been signed off as meeting the agreed criteria, CSC have not been paid the charges they would have expected to receive.
Bacon: Does committing the NHS in this way to payment for systems that may not be delivered represent good value, given the repeated failure of iSoft to deliver a working product to deadline?
In these circumstances, why is the NHS still contractually bound to minimum volumes?
Connelly: The removal of exclusivity, and thus the volume commitment, is an option available to the Department where the contractor commits a breach of contract which cannot be remedied in an acceptable way or timescale. This is, correctly, a high threshold but is an option under constant and current review.
Bacon: I am also concerned that it appears that possible providers for £100m community and child health systems at 13 Trusts in the South of England, under the Additional Supply Capability and Capacity ASCC) framework, have been deterred from bidding because of what they have called “onerous” requirements and “untenable terms and conditions”, which bias the process in favour of existing lead suppliers under the NPfIT.
Is it correct that suppliers able to deploy systems at a fraction of the cost of the LSP solutions have not been given a fair opportunity to compete?
Connelly: I find the suggestion that the Additional Supply Capability and Capacity (ASCC) Framework makes ‘onerous’ demands on potential suppliers, and includes ‘untenable terms and conditions’ particularly surprising. It may be helpful if I explain some basic background here.
Under the procurement being run via the ASCC framework more than 60 companies were, in 2008, appointed and so agreed to be bound by its commercial terms and conditions. The companies appointed ranged across small, medium and large enterprises. The ASCC framework’s commercial terms and conditions are based on the Office of Government Commerce’s (OGC’s) ICT Services Model Agreement and Guidance. The OGC is part of the Efficiency and Reform Group in the Cabinet Office. The Model Agreement and Guidance is recommended best practice.
The business requirements for the Community and Child Health procurement are those agreed jointly by the Trusts taking part. The suppliers who successfully demonstrated their capacity and capability to deliver services relevant to this Community and Child Health business need, by way of being appointed to the ASCC framework, were invited to put forward bids. The suppliers selected will be those who offer the best combination of fit with the Trusts’ business needs and the prices they offer. If a supplier cannot meet the needs of the NHS, they will not be selected.
Bacon: I would also appreciate any observations you have on how the costs of the iSoft Lorenzo, Rio and Cerner Millennium systems compare with the indicative costs submitted by suppliers as part of the ASCC market testing for acute and community systems (taking into account, of course, the costs of the various key milestone payments).
As you know, the National Audit Office is now beginning a further urgent inquiry into developments in the NPfIT, and in particular of the awarding of former Fujitsu sites to BT.
I would suggest that this inquiry will review a great deal of evidence that is relevant to the question of whether proposed contract renegotiations with BT and CSC really do represent good value to the NHS and taxpayers.
I would therefore ask for your undertaking that no further steps will be taken to conclude further agreements with either company until the NAO has reported and the Public Accounts Committee has had time to consider its findings.
It would seem to me that to enter into new irrevocable commitments in advance of this report would be irresponsible and potentially a breach of the Accounting officer’s responsibilities.
Connelly: We are actively in the evaluation phase of the procurement for Community and Child Health Trusts and, subject to final approvals, about to start the procurement process in the acute care settings. It would be inappropriate and potentially prejudicial to that process to release sensitive information relating to supplier pricing.
However, in terms of value for money, once each of the procurements is complete and full costs are known an investment case will be made for each which will be subject to both Department of Health and HM Treasury reviews specifically on value for money. As you suggest, the NAO may [sic] due course take a view about this and other aspects of our procurement programme.
Bacon: I would also ask you to undertake a specific review of the costs of running the Cerner Millennium system over a traditional PAS in Taunton; and the costs of running Lorenzo over a normal PAS in Morecambe Bay.
Connelly: I am happy to consider your request to undertake a specific review of the costs of running the respective systems, and I will discuss further with colleagues in the NHS locally.
Bacon: Please could you send me copies of any value for money reports which compare the cost and value of NPfIT systems over traditional deployments. Please include N3 Broadband and PACS as well as acute, mental health and community systems.
We have provided copious information on these matters to the NAO in the course of their latest, and previous value for money studies.
I do not think it is appropriate or necessary to provide this information separately to yourself (albeit a member of the Public Accounts Committee), particularly as it contains commercial and confidential information, whose public disclosure may be prejudicial to the Department’s ongoing negotiations
Please could you supply a copy of the indicative pricing submitted for ASCC - for acute and community trusts.
I do not feel able to provide the information you have requested, for the reasons explained above.
Please could you supply a more detailed breakdown of the 7 southern sites, in particular: the Fujitsu cost of each deployment; the Fujitsu cost of the interim running between the two contracts; the cost of the transfer, the running costs till the end of the contract; and the anticipated annual running costs payable by Trusts after the contract ends
The contracted costs of the Release 0 Cerner Millennium (the initial phase of deliver) deployment under the Fujitsu contract have been provided to the NAO
The Fujitsu cost of the interim running between the two contracts. The total value of the Short Form Agreement with Fujitsu was £76.7m. The Authority is seeking to recover some of its - and other costs - from Fujitsu via formal arbitration proceedings.
The cost of the transfer. The total cost of BT data centre commissioning and migration of the 7 remaining [Fujitsu] Live Sites to the BT data centre was £73.84m.
The running costs till the end of the contract. The total Services Charges payable to BT for acutes between contract signature (March 2009) and the Contract end (October 2015) is £127.15m.
The anticipated annual running costs payable by Trusts after the contract ends. At this stage I cannot give an accurate answer to this question. The charges payable by an individual trust cost could vary greatly depending on the size of the trust and the scope of the services, where they procure those services from etc. The LSP contracts do provide for a 1 year extension and for a further period of up to 2 years extension during the Termination Period while Services are being transferred.
Bacon’s reply to Connelly
Bacon replied to Connelly’s letter on 2 March 2011. This was Bacon’s letter in full. I have added sub-headings.
Ms Christine Connelly
Chief Information Officer for Health
Department of Health
London SW1A 2NS
2nd March 2011
Dear Ms Connelly
CONTRACT NEGOTIATION IN THE NATIONAL PROGRAMME FOR IT IN THE NHS
Thank you very much for your letter of 24th January 2011.
In my original letter to you of 13th January 2011, I asked you specifically whether the contract renegotiations with BT and Cerner represented best value for taxpayers’ money. Although I am grateful for your detailed reply, I do not see any direct answer to that question, which seems to me absolutely fundamental.
Unless the question can be answered clearly in the affirmative, I can see no justification for concluding further agreements with these companies, at least until after the NAO has reported and the Public Accounts Committee has had time to consider its findings. I am disappointed that you have not given such a commitment and I would ask you again to consider doing so.
I would also like to address some of the specific points in your letter.
My point about the Cerner product in London was that originally planned deployments of GP systems, a London Ambulance solution and a number of systems in acute trusts were included in the original contract price of £1.2 billion and are not now being delivered.
I note that you do not comment on my observation that only about half the original contract will be delivered, but that payments have been reduced by only £112 million. I cannot see how this can be value for money to taxpayers and I note that you do not claim that it is.
You say that I am “misinformed” about the level of Cerner functionality, which you say “has not been reduced”. I must refer you to page 23 of the June 2006 NAO report in which it is stated that Cerner Millennium was being supplied in four releases, covering the same functionality as the original five IDX releases.
This included a Patient Administration System, a GP system, a web based referrer, a pharmacy system, and systems for acute departments, emergency care and mental health, all to be delivered as an integrated Cerner solution. Much of this has now been dropped, or replaced with cheaper departmental systems such as RiO, but I cannot see that taxpayers have seen an appropriate consequent saving.
Broken promises on Lorenzo - will anything change?
You acknowledge that the history of delivery of Lorenzo is “very unsatisfactory”, a point on which we can certainly agree. My additional point is that there is no good reason to think that the long record of unmet promises, missed deadlines and in some cases outright misrepresentation is about to change.
Therefore, I can see no good reason for the Department to persist with supporting the deployment of the Lorenzo system through the contract with CSC. I note that the Chief Executive of the NHS and the President of CSC were scheduled to meet to discuss this matter and I sincerely hope that this does not simply result in further unreliable promises and consequent failures.
RiO price ten times higher under the NPfIT?
In relation to the price of the RiO product, I would observe that the systems supplied directly by CSE Servelec seem to be functioning in practice as well as the LSP version, so I cannot see how there is a good reason for the price of the LSP versions to be around ten times as high.
As you say, this is likely to be considered by the NAO. In the interim, I should be happy to supply a list of non-NPfIT RiO users if that would assist you in making the comparison. My concern about value for money to taxpayers is of course not related to whether the decision was made entirely by your Department or with some involvement by the Treasury.
Since my letter to you there have been further reports that raise doubts over the effectiveness of LSP systems, including:
- serious under-use of Lorenzo at NHS Bury;
- a failure to go live with Lorenzo at Pennine Care NHS Trust; and
- a call for compensation by North Bristol NHS Trust for late delivery of a PAS system.
Are BT and CSC NPfIT contracts a burden on taxpayers?
These reports simply add to my concern that you seem to remain committed to the delivery of systems through LSPs that have been shown to be unreliable, subject to serious delays and, even after contract renegotiations, unreasonably expensive.
Quite apart from the burden this places on taxpayers at a time of exception stringency in public spending, it is also surely not in the interests of the NHS and its patients.
I do sincerely hope that you will now feel able to give a commitment not to enter into revised contracts until the NAO has reported and the PAC has considered the matter.
MP for South Norfolk
Member of the Public Accounts Committee