Two computer companies are squaring up for a fight over a lucrative deal to provide software for the nationwide extension of NHS Direct, the nurse-led telephone advice service.

Presently, 13 NHS Direct pilots are up and running, using three separate software systems.

But the Department of Health plans to specify a single system to be used for the national service, due next year.

A DoH spokesperson said a national procurement exercise later this year will take note of the experiences gained in the pilot schemes. A nationwide deal will be worth millions of pounds and is being eyed by McKesson HBOC, whose Cetramax system is used in five of the schemes, and Plain Software, whose TAS system is used in six.

A third system, Personal Health Advisor, from Access Health, now also owned by McKesson HBOC, is being used in two pilot sites.

Until recently, UK company Plain Software was seen as too small to be in the running for a nationwide deal, but it has just signed a deal with the healthcare arm of Hewlett-Packard.

There are major differences between the way the HBOC and Plain systems work and it is not clear how much it would cost for those already using the systems to switch from one to the other.

Jane Coster, director of NHS Direct in West Yorkshire, which uses the HBOC system, warned that users would need time to retrain staff to use any new system.

James Munro, clinical lecturer in the medical research unit at Sheffield University, who published an interim evaluation of NHS Direct pilots in March, said more research was needed to establish what effect differences in software had made to the overall performance of the pilot schemes.

The TAS system is less structured than the HBOC system. It enables a nurse to listen to a patient's symptoms and then use the system to document and support their decision. This means that it relies more on nurses' own experience.

The HBOC system uses a set list of questions and recommendations, but claims to be less likely to miss unusual clinical symptoms and results in fewer patients being sent to accident and emergency departments unnecessarily.

Dr Munro points out that even with the same software, similar calls could result in different outcomes, depending on the local circumstances.

In Cornwall, for example, where it might take more than an hour to get to the nearest A&E department, a nurse might recommend a different course of action than in Manchester, where A&E is no more than 15 minutes away.

'It is not yet clear whether the differences in software are more or less important than the differences between users,' he said.