Big computer projects are risky. Studies show that anywhere between two-thirds and 90% of IT projects either fail or run late and go well over budget. Only a small minority are unqualified successes. The larger the project, the more likely the risk of failure.
Few organisations are larger than the NHS, which today unveiled a £500m strategy to modernise its IT by 2005. The strategy, snappily entitled Building the Information Core - Implementing the NHS Plan promises that by 2005 we will have "a vibrant networked NHS".
A blueprint for disaster? On the surface, the omens are not good.
Apart from the size of the NHS, not to mention its past record in IT, there are two reasons to worry.
First, unlike other big organisations such as supermarkets or banks, the NHS's head office cannot demand that every hospital or GP surgery install a uniform set of IT systems.
Every NHS organisation, even down to individual GPs, has its own IT strategy and mixture of "legacy" computers which may or may not talk to each other.
Second, healthcare everywhere is more difficult than other industries to computerise. Medical language is far more complicated and subtle to digitise than sales figures or stock inventories, and patients' bedsides are not good places to install computer terminals.
The authors of the IT strategy know the risks, which presumably explains its low-key launch. But there are reasons for optimism. By the standards of Labour's grandiose scheme's to deliver e-public services, the strategy has been carefully thought out.
It also builds on, rather than overturning, previous initiatives. For example, the target of all acute hospitals having electronic patient records by 2005 was set in the previous strategy, Information for Health, published in 1998.
Some targets have even been scaled back: in 1998, ministers were talking confidently of connecting all GPs to NHSnet by March 2000: that date has since slipped to 2002.
Technically, the strategy does not look particularly ambitious. One headline target, 2005 for electronically booked appointments, would appear painfully slow in other industries.
Patients might wonder why the wait - after all, this is what happens every time a travel agency makes an airline reservation.
In practice, it's all a bit more difficult. Even when the technology is sorted out, there are complex political questions. GPs are supposed to "request" appointments from specialists, who decide to allocate them according to available resources, not order them like cans of beans.
There are also uncertainties surrounding the two tiers of electronic records envisaged in the strategy.
By 2005, all acute trusts to hold electronic patient-records (EPR). Such "organisational" records are difficult enough to create - only a handful of hospitals in the world have them, and about one-tenth of NHS trusts have moved significantly down this path. (Their use by GPs became legal only in October last year.)
However by 2005, the strategy envisages the NHS piloting a second tier electronic health record (EHR), a cradle-to-grave dossier of essential information available seamlessly to health professionals, the patients themselves, but not to insurance companies, police officers or journalists.
Arguments about the security of electronic records have hamstrung healthcare IT for a decade, and are not about to go away. The strategy revives one possible solution, the idea of issuing patients "smart cards" as keys to their records.
But here the strategy hedges its bets, saying that it awaits the outcome of two separate studies due to report by March.
There are other reasons for caution. While the new money for IT is welcome, it will do no good if NHS organisations are not allowed to spend it.
Today, procurements usually take two years or more, to the despair of IT companies, many of whom have left the market. The strategy's solution is to draw up national lists of approved systems, from which hospitals can pick without unnecessary bureaucracy.
The snag is that such lists take years to draw up, with IT firms fighting like ferrets in a sack to get onto them.
In the meantime, technology moves on and the targets loom closer. There are no quick fixes.
