A LONDON hospital last night ruled out 'death by computer' after one of its doctors threatened to use a machine to predict which patients would live - and were cost- effective to treat - and which would die.

A spokesman for the Guy's and St Thomas's Hospital Trust said that Dr David Bihari, a consultant in the intensive care unit (ICU), did not have authority to use the computer in clinical decision- making. Dr Bihari's colleagues accused him of scare-mongering to make political points about the NHS before moving to Australia.

Dr Bihari said that budget cuts would force him to make clinical decisions with the aid of the US 'Domesday' computer. This flashes up a coffin symbol for any patient unlikely to survive 90 days. He said he would stop treatment for some patients although research suggests the computer may be wrong in one in 20 cases.

A spokesman for the Guy's and St Thomas's Hospital Trust said the computer was being used for audit only, to assess the performance of the ICU.

Dr Bihari had not asked the trust's ethics committee for clearance to use it to make decisions about patient care. Dr Bihari, an outspoken critic of the health service changes and particularly the merger of Guy's and St Thomas's, made his claims on a BBC 1 Television programme, Here and Now, last night. He was unavailable for comment yesterday but consultant colleagues were angered by his 'alarmist' statements. One said: 'He is a well-known self-publicist and that is what this is all about.'

Dr Gilbert Park, director of ICU at Addenbrooke's Hospital, Cambridge, said that the RIP computer (Riyadh Intensive Care Programme) had been developed for audit purposes only. 'These computers have been in use in the US for 10 years and we have had one in Addenbrooke's for five years. There is not a lot new about them really,' he added.

Dr Kathy Rowan, director of Audit and Research for the Intensive Care Society, said that she was unaware of any published research to back up Dr Bihari's claim that the computer had a 95 to 96 per cent success rate in predicting survival. It uses the patient's medical history and current condition and draws on the outcome of hundreds of other cases to reach a decision.

Dr Bihari told Here and Now that he had to cut his pounds 3.3m budget by between 3 and 7 per cent. Asked if the computer would have a direct input into the lives and deaths of patients, he said: 'Yes, it will help me and my staff identify those patients who should have their treatment stopped. All I know is that I am trying to provide a service and given a certain amount of money . . . and if it's not enough I have to act accordingly.'

Dr Richard Nicholson, editor of the Bulletin of Medical Ethics, said that a computer could not replace a doctor's clinical input but would aid it. 'The computer doesn't switch off the ventilator, the doctor does. This is no different than other methods of predicting survival used for years.'