The NHS database is in a critical condition

It started on Tony Blair's sofa. No longer should the NHS rely on a hotch-potch of elderly computer systems, they told him. The days when patients had to repeat themselves every time they saw a new doctor would be over. They would build a new NHS computer system, a £6 billion grand-projet, and at its heart would lie a vast database that would make researchers salivate. They would build a new NHS. They would revolutionise everything.

And in twenty minutes, so rumour has it, the deal was done. Blair had bought it.

It hasn't quite worked that way. The care record system, the bundle of patient information on which everything else would be built, was meant to arrive in 2004; five years later it is still five years off. In 2007 Accenture, one of the four 'local service providers' (LSP) hired to deliver it, abandoned the project; last year Fujitsu followed it out of the door. After seven years, most hospitals still haven't got their new computer system, and those that have are struggling to make it work. Most striking of all, the price of the whole project is now estimated to exceed £12 billion.

This week, the Tories have promised to save the day. In a report published on Monday, the party pledged a comprehensive overhaul of the programme. It promised to renegotiate the mammoth LSP contracts that had failed to deliver. It would allow hospitals to buy whatever software they wanted, providing it met certain standards. And it would scrap that faintly sinister national database, and replace it with alternatives as 'small and localised as possible'.

Designing a patient record for everyone from GPs to brain surgeons is impossible

The party pitched its plans as a radical shift. And to an extent, they've been greeted as such, largely thanks to a few under-developed ideas about allowing Microsoft or Google to host patients' records.

Actually, though, the report is largely a statement of the bleeding obvious, and closely echo what those who actually work with NHS computers have been saying for years.

Designing a single patient record for everyone from GPs to brain surgeons is impossible, they say. And anyway, grand-projets move far too slowly to respond to changing needs. Much better