16 June 2006
Connecting for Health – the NHS's huge National Programme for IT – has been dogged by 'scope creep' and core 'spine' problems. But as its chief, Richard Granger, tells PF, plenty has gone right, too. Michael Cross reports
Amid last week's headlines about the English NHS's budget deficit, one health service agency basked in a strange distinction. NHS Connecting for Health, set up to implement the National Programme for IT, still has in hand more than a billion pounds out of the £2.3bn it was allocated for three years in the 2002 Spending Review.
'We have well underspent on those numbers,' Richard Granger, director general for NHS IT and chief executive of the programme, told Public Finance in an exclusive interview. Although Granger, recruited from the private sector four years ago, could not provide a breakdown, published figures show that Connecting for Health – one of the world's most ambitious civil IT programmes – had spent £1,024m by January 31 in running costs and payments to suppliers.
Meanwhile, industry figures released this week show that overall frontline NHS investment in IT, vital to the programme's success, has remained almost static for four years.
The underspend rebuts political and media assumptions that Connecting for Health, which is expected to cost some £20bn over ten years, is massively exceeding its budget. However, the revelations add to a sense of crisis around it. At the end of May, the health minister responsible for IT, Lord Warner of Brockley, admitted that one reason for the underspend was that the centrepiece system was more than two years behind schedule.
The crisis is now likely to come to a head with the publication of a long-delayed National Audit Office report next week and a hearing at the Commons Public Accounts Committee on June 26. Among the problems that will be debated is the failure of suppliers to deliver systems on time, particularly specialist medical software. Isoft, for example, which was contracted to supply clinical software to three of the five NHS IT regions, last week announced it was laying off 15% of its UK workforce and that it was in talks with the NHS about rescheduling delivery dates.
Some NHS trusts have also been reluctant to invest locally in infrastructure to handle the new electronic patient record systems. Granger confirms reports by the IT industry association Intellect that overall NHS spending on IT is well below the 4% target set by Sir Derek Wanless's 2002 review of NHS finance, Securing our future health: taking a long-term view. IT expenditure is 'inching up... but it is not going up towards 4%', Granger says.
But he defends Connecting for Health against media attacks. As a former management consultant with extensive experience of government IT projects, he particularly objects to the programme being lumped in with the long list of Whitehall disasters.
While 'there are elements that are late', he says, overall 'we've done it roughly to timetable'. Over the past two years, the programme's achievements include installing a 'core infrastructure' – connecting NHS organisations securely and enabling the beginning of electronic appointment booking and prescribing.
The programme team has also installed a host of systems that were never in the original specification – notably electronic X-rays, now available in half the hospitals in England. As to the future, Granger is confident that the government's NHS reforms will increase the importance of national IT systems. 'Nothing that we set out to do has become redundant.'
Although some trusts are having difficulties implementing some systems, he says, the story varies widely from location to location. He implies that local management bears some responsibility. 'We know there's a correlation between the maturity of the organisation and the lack of disruption as they put these things in.' One of the programme's priorities this year is to iron out the discrepancies in the experience NHS organisations have with installing new systems.
Granger also refers to hostility to the programme by some professional bodies and a negative media attitude. 'When things are going well, you get only silence.'
The NAO report will give the first independent, authoritative, verdict of whether the programme is running to time and represents value for money. Granger would not comment on early drafts. 'We are working to ensure the report is based on evidence and is balanced,' he says. Would that imply that the first drafts were unbalanced and lacking an evidence base? He smiles enigmatically.
Any fair picture of whether the IT programme is on track must look back at what it set out to achieve. The NHS has had an IT strategy since 1992, originally to support the flow of contracting data required by the Conservative government's internal market. In 1998, a new strategy, Information for Health, set deadlines for medical records to be computerised. Progress was patchy – mainly because IT was considered a low priority for trust chief executives. Much of the extra money allocated to support Information for Health – £214m between 1999 and 2001 – was spent elsewhere.
In April 2002, the Wanless Review recommended that the increase in NHS budgets should be based on modernised IT. Describing existing systems as 'piecemeal and poorly integrated', Wanless proposed that the NHS double to 4% the proportion of its budget invested in IT, to bring it closer into line with health care providers in the US (6%). The review also proposed that the NHS should buy IT on a national scale, rather than leaving negotiations to individual trusts or surgeries.
Wanless was pushing at an open door. A decision to transform the NHS's IT had been taken at a seminar in Downing Street in February 2002. Attendees are understood to have included Sir John Pattison, then the Department of Health's head of research and development; David Bennett, then a partner with McKinsey (and now head of the Number 10 Policy Unit); and representatives of two IT suppliers. In summer 2002, Pattison published an outline of how the programme would be run, and what it would achieve.
The system as depicted had three pillars: electronic prescriptions, electronic appointment booking and a lifelong health records service, based on a 'standard system specification'. These components were to be in place by the end of the programme's second phase, in December 2007. Systems would be run by contractors appointed by strategic health authorities rather than trusts.
In October 2002, Granger took up his post. The following month, the Spending Review announced that the National Programme for IT would receive £2.3bn in central funding over the following three years.
Granger's personal style and approach to the job set the programme apart from its predecessors. Aided by a small team recruited largely from outside the NHS (including consultancy Kellogg Brown & Root), he introduced a unique set of procurements under principles that stressed speed, competition and payment only on delivery of working systems. Three years on, Granger vigorously defends those principles.
The procurements were deliberately packaged to create very large ten-year contracts. England's 28 strategic health authorities were grouped into five 'clusters', with boundaries drawn up to be coterminous with the Regional Offices for England rather than traditional NHS boundaries. In each cluster, a single prime contractor was to take responsibility for updating systems. There were also two national contracts, for an electronic booking service and the 'spine' of the central health records system, now known as the care records service.
Remarkably, all seven contracts were let within a year (half the time typically taken by an individual hospital trust to pick a core IT system). The process kept at least two bidders in every race until the very last moment rather than negotiating with a 'preferred bidder'. The contract values added up to £6bn, which Granger says represents a fraction of the price of systems procured through previous methods. Because of the size of the contracts and the scale of risk, the procurements attracted bids from firms that were either new to the NHS or had withdrawn from the market when it fragmented and stagnated in the 1990s.
They included Accenture, which won two cluster contracts, and CSC and Fujitsu, which each won one. BT, which was already heavily involved in the NHS, won the cluster contract for London and the highly strategic £600m deal to run the 'spine'. Schlumberger Sema (now Atos Origin) won a £65m contract to create electronic booking software.
To develop the actual software needed to make medical records available anywhere in England, the service providers picked two specialist subcontractors: Isoft for England north of London and the Severn, and IDX, a US-based supplier, for London and the south.
The programme's implementation phase began in 2004. The first major deadline was for all GPs to be booking appointments electronically by the end of 2005. It was missed.
One problem was 'scope creep'. Unforeseen challenges loaded on the programme included the need to procure new broadband network capacity, a new NHS-wide e-mail system to replace one run by a contractor terminated for underperformance (EDS) and a system to support payments to GPs under the new contract.
In 2004, the programme also negotiated contracts with suppliers of picture archiving and communications systems (Pacs) to encourage the early adoption of electronic viewing of X-rays and other images. Pacs has a strong business case, in that electronic imaging removes the need for expensive X-ray film. Importantly, the cash savings fall to the institution making the investment, which is not always the case with IT spending in health care.
A more serious scope creep was the electronic booking service's metamorphosis into the Choose and Book system, supposed to underpin the 'choice-based' NHS reforms. While the software was delivered on time, the care records spine, upon which Choose and Book relies for patients' demographic information, ran late and in any case the policy and management changes required to implement Patient Choice had not been put in place. For example, the Connecting for Health team found itself having to compile a directory of NHS services.
As a result, the programme publicly missed its target of universal electronic booking by the end of 2005. About 8% of appointment bookings by GPs are now made electronically but usage of the system is growing exponentially, according to Simon Eccles, one of the programme's clinical leads.
Electronic prescribing likewise turned out to be a more complex policy headache than originally expected. In 2002, transmitting prescriptions electronically from doctor to pharmacy to Prescription Pricing Authority was seen as an early win for the programme, because of the obvious waste and risk involved in the dual retyping of 350 million prescriptions a year. However, the project had to be put on ice while the Office
of Fair Trading investigated the future of the retail pharmacy industry, including the role of supermarkets. Although the system is now being implemented – still running in parallel with paper prescriptions – unresolved complications include confidentiality issues involved in giving pharmacists access to electronic health records and the question of how to deal with controlled drugs. Connecting for Health says that more than 1.2 million prescriptions have been issued electronically to date – a tiny fraction of the 350 million issued each year.
By far the most difficult implementation issues facing the programme, however, relate to the design and function of the care records spine, and the practicalities of connecting it to IT systems used on the front line. These are the areas in which the unique features of Connecting for Health – its national, mandatory character and ruthless commercial terms – cause the biggest upheaval.
The care records service, due to go live by the end of 2007, will create electronic health records by combining central data about patients with summaries of medical treatments drawn from hospitals' and GPs' electronic patient records (and, eventually, social care case files). Its underlying system is the spine, whose production version went live in January.
This carries demographic data to support Choose and Book, as well as some electronic prescriptions and whole patient records when patients transfer between GPs. Early next year, Granger says, pilots will begin of live clinical information on the spine. This is the part of the programme that is running at least two years late.
One problem is technical. Standardising systems to ones accredited for connection to the spine is a massive task. In many cases, it requires NHS staff to adopt systems that, in the short term at least, represent a backward step from those already installed. Because of delays in the new software, prime contractors have been forced to rely on a mix of 'interim solutions' to stay in business. This has had consequences for their cash flow: in March, Accenture's annual accounts reported that it was setting aside $450m to compensate for losses on its NHS contracts. This was reportedly due to delays with Isoft's new system.
The other clinical software supplier, IDX, appears to be having similar delivery problems to Isoft. The Southern contractor Fujitsu last year replaced IDX with another US firm, Cerner. Then last month, BT, prime contractor in London, said that to give IDX more time, it had struck an extension deal with a previous supplier McKesson, which IDX was supposed to replace.
Another problem is getting agreement from doctors about what the summary record should contain. This is highly controversial. Granger expresses frustration because, as far as the programme was concerned, this was settled two years ago. However, the bulk of the medical profession is only just addressing the question.
'The reality of having a working infrastructure hooked up to tens of thousands of organisations is finally entering the consciousness of people.' One unresolved question is over confidentiality – whether patients should have to opt in to the system, or be assumed to consent unless they opt out.
Another area of evident frustration for Granger is the apparently straightforward one of transferring records electronically between GPs' computer systems. With 3.5 million people changing doctors every year, this is a substantial flow of information. Early in the programme, Granger identified this as a priority, but so far the only GP-GP electronic transfers that are taking place are where the doctors have identical software. Details on how it will be achieved between different suppliers are still being negotiated with the BMA's general practice committee.
Whatever the NAO's verdict will be this month, the NHS National Programme for IT seems to need a new impetus four years after its conception. Technically, the most difficult parts of computerising health records lie ahead – especially that of combining information from GPs, hospitals and (eventually) community care. 'There are no software products out there at the moment that can do that,' Granger says. He speaks of a 'mixed economy' of systems, rather than the original vision of an all-purpose clinical information system. However, this model makes the centrally run spine more, not less, crucial, to connect everything together.
The programme also desperately needs to build support in the NHS. One idea expected to be announced shortly is to create an exemplar region combining all the advances being tested individually around the country. 'One community where everything is working,' says Granger.
A 'refresh' of the programme's approach, expected shortly, will also clarify the relationship to the white paper Our health, our care, our say, answering criticism that the centrist national programme is ill-aligned with the government's NHS reforms.
Granger says there is no such problem. The whole thrust of the reforms is to offer care in more than one location. 'What's going to make that work? Having the right information available to support people.' As for the mixed economy: 'The programme is for the profit-making sector as much as anyone else. We are going to provide the cohesion through which the pluralistic supply operation will function.'
But if IT is going to be the defining feature of the NHS, users and patients must have confidence that it will work. Granger says this is lacking – he blames hostile and uninformed press coverage for obstructing change – 'an environment in which the realities of putting in large complex changes are being misrepresented'.
He professes himself bewildered by the attention. 'I thought that saving thousands of lives and improving the effectiveness of hundreds of institutions would be a largely uncontroversial activity. I believe that's what's called naive!'
Michael Cross is a freelance journalist and specialist writer on health care informatics
PFjun2006
