Only a week ago a deal aimed at rescuing the NHS's National Programme for IT in the south of England seemed imminent.
Officials and Fujitsu had spent nearly a year negotiating changes to a 10-year contract which was worth £896m when signed in January 2004 and had increased for various reasons to a little over £1bn.
A deal in principle was agreed, ready for signing by David Nicholson, the Chief Executive of the NHS who is also the predominant senior responsible owner of the £12.7bn National Programme for IT [NPfIT].
But at what one NHS official said was the "59th minute of the eleventh hour" Fujitsu informed Nicholson that it was withdrawing from the negotiations. In response, the NHS has said it is terminating Fujitsu's original contract. There had been a threat by the NHS of terminating the contract even during the "contract re-set" negotiations. But still the withdrawal of Fujitsu has taken many in the NHS by surprise.
One executive involved in the NPfIT said:
"Fujitsu's withdrawal is not a real surprise but the manner in which it happened was. Up until the end of last week the signs were that agreement had been reached. Things were looking positive on a reasonable implementation timetable."
A spokesman for NHS Connecting for Health, a part of the Department of Health, said today [29 May 2008]:
"Regrettably and despite best efforts by all parties, it has not been possible to reach an agreement on the core Fujitsu contract that is acceptable to all parties. The NHS will therefore end the contract early by issuing a termination notice.
"Work has started immediately on planning the necessary arrangements."
How did a proposed deal collapse?
It's thought likely that the Japanese board of Fujitsu intervened. Some NHS officials believe that Fujitsu's board in Japan decided that its potential losses on its NPfIT work, as one of three local service providers to the NHS, could be much greater than its directors had thought at first.
According to figures released under the Freedom of Information Act, Fujitsu received in one year alone, 2006/7, about £264m. Some of this has been for supplying digital x-ray equipment - picture archiving and communication systems [PACS] - which has been a successful part of the NPfIT. PACS was added to the programme two years after the scheme was announced. And some of the money paid to Fujitsu was advance funding, refundable in cash or the equivalent in services.
Fujitsu's receipts under the NPfIT are much smaller than those of the two other local service providers, CSC and BT. In 2006/7 CSC received nearly £334m. BT received in the same year nearly £393m. BT and CSC have much more business under the NPfIT than Fujitsu. But as at 31 March 2008 Fujitsu had the most deployments for which no payment had been made because trusts had not signed off acceptance - 56 deployments in the south against eight in London and 48 in the rest of England.
Fujitsu made the point to the National Audit Office that trusts had at times withheld payment even when systems were working until all non-contractual changes they were seeking had been implemented.
The supplier's losses on its NPfIT work are estimated at about £340m. Its original £896m contract had already risen to about £1.07bn because of a change of the main software supplier from IDX to Cerner and additional work.
During Fujitsu's re-negotiation with the NHS there was discussion of the contract price rising to more than £1.4bn but this was negotiated down to about £1.15bn, a sum agreed by the NHS before the withdrawal.
Now that the NHS has said it is terminating Fujitsu's contract, the supplier is expected to have 90 days, or more if specifically negotiated, to run down its work on the NPfIT. Few in the NHS will be surprised if Fujitsu receives some further payments, at least for giving support to "early adopters" of the Cerner "Millennium" Care Records Service. Fujitsu's contract was to deliver Millennium at NHS sites across the south of England.
A smooth run-down is expected. The last thing the government wants is litigation with Fujitsu.
But all that the NHS had been relieved to negotiate in the contract re-set has evaporated.
Negotiators had thought they'd secured a deal which would provide access, for the first time, to the full Cerner Millennium product - not a reduced one-size-fits-all version which restricted how much the software and NHS data within it could be manipulated by individual trusts. Two NHS trusts where Cerner's product has been successful are Homerton and Newham in London. But they bought the system directly from Cerner, outside the NPfIT.
An NHS IT executive said:
"Homerton and Newham had the full product. This was what contract re-set would have given the NHS - the full product. A lot of the vagaries that the early adopters had to deal could have disappeared at other trusts because the full product could be configured to your requirements. You would get your own environment. A great deal of what the NHS wanted was in the contract re-set."
So what happens now?
Few trusts are prepared for Fujitsu's withdrawal.
The board of South Downs Health NHS Trust has been told that because of the uncertainties over the contract re-set "all Sussex trusts are looking at interim care record solutions pending moves to a viable integrated care records which now, realistically, looks several years away".
A Sussex-wide review of options took place in March 2008. But time is pressing. Like many other trusts South Downs needs to replace its systems. Its PiMS patient information system is running is no longer supported by the supplier iSoft and it is on an unsupported Oracle database and the Windows 2000 operating system.
The NHS official said:
"We're into a period of turmoil. There's the exit arrangements with Fujitsu to manage especially for the early-adopter sites and there will probably be some months of discussion about what we should do. There is no consensus within the NHS on that." So NHS trust directors who have waiting months for advice on what to do next face uncertainty piled on uncertainty.
There are several options. One is for the government to give trusts the freedom to buy under Connecting for Health's "ASCC" - Additional Supply Capability and Capacity - framework which has recently come into force.
Unless an NHS trust has foundation status it has a duty to buy the NPfIT Care Records Service software from a local service provider. But if this mandate were dissolved it would allow trusts to use the ASCC framework contracts to buy directly from Cerner, or buy another hospital system from one of a small group of suppliers of patient administration systems. This would eliminate the local service provider as a middle-man.
It's unclear, though, whether the chosen system would be free of charge to trusts - funded centrally and not from local budgets - as was Cerner when supplied by Fujitsu. Most IT executives in the NHS are expected to favour this option, particularly if the software is funded centrally irrespective of what ASCC choice the trust makes.
A second option would be for Fujitsu's work to be passed to BT, the local service provider for London. But deployments of Cerner in London have been spasmodic and not without difficulty. BT's main installations of Cerner at Barts and The London NHS Trust, Queen Mary's Sidcup and Barnet and Chase Farm Hospitals NHS Trust have run into some serious problems, although all three have overcome some sizeable obstacles.
Some NHS executives believe that that giving Fujitsu's contract to BT would mean more of the same. There have been delays of four years in the main part of the NPfIT, the Care Records Service, according to the National Audit Office.
Fujitsu's work could be passed to CSC, the local service provider for England outside London and the south, which is due to supply to NHS trusts the "Lorenzo" Care Records Service. But though the NPfIT contracts are five years old, the national, strategic version of Lorenzo has yet to be installed in any trusts.
A fourth option would be for CSC and BT to take over jointly Fujitsu's contract. This would give smaller trusts in the south expertise on tap from the installers of Cerner or Lorenzo. But it would mean the NHS is still restricted to two main software products that have yet to prove they merit a duopoly.
Whatever the options where does the withdrawal of Fujitsu leave the NPfIT?
NHS officials say that now, five years into the National Programme, they would have expected to be three years into a roll-out.
"Actually we are barely one year down the path of a go-live programme and it has taken five years to get a small number of small trusts in the south to go live [with Cerner]. The best go-live anywhere - and it was a very limited version - was probably Taunton and even there they have had loads of problems. Can you call all that a success?"
NHS staff believe that Cerner can be made to work across the UK. It's a successful product in the US and elsewhere. But there's also a concern in the NHS the Millennium system is a client-sever system rooted in the 1990s. NHS staff need extensive training. It's not as intuitive as, say, a Web-based online banking system which needs no training.
There's also concern among some officials that the NPfIT is dated in concept as well as practice. Since the programme was announced trusts are subject to competition for patients from private companies and even within the health service, particularly foundation trusts. They want IT to give them a competitive edge, which makes them less inclined to favour systems chosen for them centrally.
The executive working on the NPfIT said:
"Had there been a more competitive market we'd have had something better sooner. If you try to manage the market excessively you stifle innovation and risk-taking and the sorts of things that drive most markets wherever you look. You wouldn't have such pervasive use of mobile phones now if someone somewhere had said: 'Let's have a national programme for mobile phones ... and the only option was a Nokia brick'."
There's still strong support within the NHS for the idea of a reliable electronic health record to replace paper records which are lost frequently or not available when needed. But there's a poor perception of the NPfIT as a vehicle to deliver electronic records, which will be exacerbated by Fujitsu's withdrawal from the NPfIT.
One official said: "The reason the NPfIT has failed in the view of many is the original approach. The government was sold the idea that this could happen in little more than two years and that was never possible. Non-delivery has been a characteristic of the programme, as has been red tape, fudge and central committees deciding it was going to be this way. And of course that doesn't work. Each time it takes another six months to find out it's not going to work."
It will be of little comfort to the Department of Health and ministers that Computer Weekly warned them in 2002 and 2003 that the NPfIT was too ambitious to be achievable, and that the programme incorporated some of the biggest mistakes of the past. For this warning ministers and some parts of the media branded us doom-mongers.
We still hope our critics will prove us wrong. But it's six years since the NPfIT was announced. How much longer do they need?