¿ìè¶ÌÊÓƵ

Earlier this month contractors arrived at a large hospital in south
London to replace its main computers with newer models capable of running
more up-to-date software. The old computers had been installed barely a
year before. Most industries would have seen the need to replace computers
after such a short time as admission of failure. In Britain’s National Health
Service, the process towards computerisation has been so slow, tortuous
and riddled with dead ends that the replacement of yet another set of unsuitable
hardware passed almost without comment.

Computers are one element of government spending in health that almost
everybody agrees will benefit from the Conserva-tive Party’s victory in
the general election. Information technology is essential to the reforms
that the government is making to the world’s largest public health service.
But the programme has a long way to go.

The NHS, founded in 1948 to run health care in Britain, employs about
1.25 million people, including part-timers, and spends about £30
billion a year. By international standards, the NHS is efficient: it offers
free health care to the population at a cost of 5.8 per cent of GDP. The
average among developed countries is 7.6 per cent. However, the system is
straining under the steady increase in the number of patients (largely
due to the growing number of older people) and relentless increases in
the cost of medical treatment as new techniques, drugs and equipment become
available.

Business discipline

The Conservative government says it is tackling the crisis by making
the health service operate more like a private organisation (although during
the election campaign, it repeatedly denied any plans to turn the NHS into
a private organisation). The government argued that introducing competition
would improve efficiency.

Many of its reforms depend on computers. A large part of the NHS’s work
involves collecting and handling information, from lists of people in the
population to medical records (including images such as X-ray pictures)
to prescriptions, letters, staffing rosters and mountains of administrative
forms. Yet, the health service is woefully backward in the technology to
handle information by the standards of private industry.

Several technical reasons lie behind medicine’s resistance to computing.
One is that the language of medicine does not lend itself to coding in a
way that computers can handle. It abounds in synonyms: doctors may use the
words heart attack, myocardial infarction or coronary thrombosis to describe
the same event. Conversely a term such as ‘chest pains’ may cover a large
number of injuries and diseases.

Also, data needed to manage a medical case usually come from a large
number of scattered sources, such as a GP’s surgery and a nurse’s bedside
observations as well as from a hospital’s registration desk.

But perhaps the greatest obstacle of all is the widespread reluctance
of doctors and nurses to use computers, which they tend to see as a bureaucratic
distraction. ‘Most health care professionals are not keyboard literate
and don’t want to be keyboard literate,’ says Colin Gunner of the computer
company Bull, which is focusing on hospital computers.

Many of these obstacles exist across the developed world. ‘Even in the
US, health care is the last major industry to be computerised,’ says George
Giorgianni, a vice president of HBO, a leading American supplier of hospital
systems.

But Britain’s special problem is lack of investment. ‘Many forward-looking
businesses spend more than 10 to 15 per cent of their operating costs on
computer and communication systems,’ says Tim Benson, managing director
of Abies, a company that specialises in medical computers. ‘In comparison,
the level of expenditure in health care computing, less than £200
million for the NHS in 1991, runs at less than 1 per cent.’

And much early investment went to waste. In 1990, a report by the National
Audit Office concluded: ‘Until the late 1980s, development of information
systems and information technology in the NHS proceeded without clear strategies
. . . Management of computer projects was often weak, with many failures
to follow good practice, resulting in poor value for money.’

The failure to fully computerise the health service in Britain has
hindered the government’s attempts to create an ‘internal market’ in health
care which, like all markets, can work only when supplied with information.

The first systematic approach to put patients’ records on computers
began in the early 1980s, when a committee chaired by Edith Korner recommended
collecting basic information about patients. These ‘minimum data sets’
became known as Korner returns.

To cope with these data, hospitals bought large mainframe computers
to handle patient administration systems. These had been around since the
1960s: in fact, development work on the most widespread system, the Inter-Regional
Collaboration patient administration system (IRC-PAS), began in 1967.

The story of the IRC-PAS exemplifies what can go wrong when politics,
health care and IT mix. About half the country’s 196 health authorities
chose the system for their hospitals, largely because it ran on machines
made by ICL, Britain’s last maker of mainframe computers (now owned by Fujitsu
of Japan).

Although the basic PAS has seen many improvements over the years – it
no longer needs to run on computers the size of wardrobes, for example,
although in practice it often does – the system is showing its age. In
1991, the government attempted to inject ‘commercial disciplines’ into
the NHS’s computer operations by awarding Qa Business Systems the contract
to maintain PAS. Only six months after winning the contract, Qa went into
receivership, delaying vital updates to the software.

Meanwhile, in 1986, the government launched the Resource Management
Initiative. This spawned a new computer system known as the case-mix management
system. It was designed to collect information about the costs and resources
devoted to each patient, relying on data from PAS as well as computers that
manage nursing services, laboratory tests and departments such as pharmacies
and operating theatres.

But, just as the Resource Management Initiative was getting under way,
the government radically changed the structure of the health service. In
April 1991, it divided the NHS into ‘purchasers’ and ‘providers’ of health
care. It also allowed individual hospitals to become self-governing trusts,
selling their services to health authorities and private individuals.

Communicating with contractors

These reforms needed two new types of computer system. The first was
known as the Development Information System for Purchasers. It would manage
the links between health authorities and their contractors, such as hospitals
and doctors’ practices. The NHS set up groups at four health authorities
to develop the system with private companies, setting a deadline of April
1992. The fruits of this work will go on show at the National Exhibition
Centre in Birmingham next month.

In addition, hospitals under the new regime were to have an a dvanced
hospital information support system, or HISS, which puts all activities
in the hospital onto a common computer system. The government chose three
pilot sites to develop HISS technology. The first of these, at Greenwich
in London, is now coming on line.

HISS has proved costly and contentious. The government had originally
planned to put systems into 250 hospitals over five years, at a cost of
between £2 million and £3 million each. But the cost of the
projects kept growing. In the end, contracts for the pilot sites cost many
times the original estimate.

Neither of the companies that had supplied the vast bulk of British
hospital computers, ICL and the American-based SMS (using DEC machines)
decided to bid. Executives of the two companies say they were horrified
by the detailed specification produced by the NHS. It was most demanding
at the third pilot site, Nottingham. This was to be a test for yet another
project, an attempt to write a set of rules called the Common Basic Specification
which define everything that happens in a hospital.

It took two years to find a company willing to take on the job. The
American giants IBM and DEC backed out, leaving the field to Oracle, an
American software company, which had never developed an HISS before. After
intervention by the junior health minister, Stephen Dorrell, anxious to
avoid embarrassment in the run-up to the general election, the NHS signed
an £8.85 million contract for the Nottingham project on 21 February
this year.

Although Oracle is confident that the Nottingham HISS will succeed,
the rest of the industry sees it as a dead end. While agreeing that a modern
health service needs integrated computer systems, few in the industry believe
that the country can afford the investment that the systems would need.
Nor does Britain have the trained people to install them.

The NHS senior management seem to agree. They are now talking about
linking up existing computers and installing new ones where necessary rather
than replacing the entire information system.

Another approach is to have no central computer at all, but a network
of machines distributed around the hospital wherever needed. This, however,
poses its own problems. The health service suffers particularly badly from
incompatibility between makes of computer: it is not unknown for a health
authority to have machines from nine or ten different manufacturers. Another
source of confusion is the sheer number of software suppliers in the market.
A recent study by the University of York found 23 systems for managing nurses
alone. The cheapest program for this task costs around £10 000.

Complex as HISS is, it is only a step towards the electronic medical
record, the Holy Grail of medical computing. This involves a lot more than
simply handling facts and figures about each patient. Much data – X-rays
for example – come in the form of images, which consume too much processing
power for today’s computers to handle. However, Hammersmith Hospital has
just become the first hospital in London to store X-ray pictures electronically,
rather than on film. Future ‘multimedia’ patient records may consist of
files containing data, still images, sound recordings and video sequences.

Further into the future, computer companies are looking to more exotic
technologies, such as artificial intelligence to help make diagnoses and
manage patients and even ‘virtual reality’ to help doctors experience patients’
aches and pains.

However, such progress is unlikely without strong central direction.
Duncan Nichol, chief executive of the NHS, told a recent conference a fully
computerised health service will not become a reality until each individual
has a unique identification number. A European project called SHINE, Strategic
Health Informatics Networks for Europe, plans to use an identification number
in a scheme to link health systems throughout the European Community. But
this concept raises issues of privacy and the defence of civil liberties.
As one political debate over medical computing closes, another may just
be opening.

Michael Cross is a freelance journalist.