Scientists have created a new way of predicting how long terminally ill cancer patients have to live, and the guide could be used with mobile phone technology.

The new computer model guide – a project led by St George's, University of London – uses factors such as blood tests, white cell blood count, pulse rate and patient symptoms to estimate how long a patient has left. The researchers believe it can predict survival at least as well as a doctor.

Scientists said the model differs from previous scales in that it can help give a more accurate picture of whether patients might have only two weeks or two months left to live, independently of a doctor’s estimate. The scale could help families, carers and nurses make plans with cancer patients who are close to the end of their life.

The team hopes that clinicians can use the guide at patients' bedsides by using mobile phones or other hand-held devices.

The study, funded by Cancer Research UK, looked at 18 palliative care services including hospices, hospital support teams and community service and more than 1000 patients with advanced cancer who were no longer receiving treatment.

A combination of markers such as blood tests, pulse rate, weight loss, tiredness, breathlessness and white blood cell count were used to produce two versions of the scale.

Lead study author Dr Paddy Stone of St George's said: “These scales can provide valuable information for patients, carers and health professionals. It is important to remember that these results do not provide a definitive model for predicting how long someone will live, but it will give everyone concerned a clearer idea of what it is likely to happen.

“This study provides a solid starting point for improving accuracy in survival predictions which can continue to be refined and improved.”

Professor Chris Todd of the University of Manchester, another author of the study, said: “An instrument like this will also help us identify which patients could take part in studies aimed at improving the quality of life for people receiving end of life care. We are already looking at how to improve the prediction models and how to make them readily available to clinicians through, for example, iPhones and other mobile devices.”

Scientists claim that one form of the scale, which does not require a blood test, provides a prediction of survival as good as a doctor’s estimate. They say that another version using a blood test is better than a prognosis.

The model could also be adapted to a patient who may not be able to respond to questions about their health.

Mike Hobday, head of campaigns at Macmillan Cancer Support, said: “This scale could prove useful to patients, families and clinicians who are wondering whether to begin discussions around palliative care arrangements.

“All too often this conversation is left until it is too late to make arrangements while patients wait to know what their future is. Having the conversation at an earlier point, alongside ensuring a 24-hour community nursing service is in place, will vastly improve the chances of the 57 per cent of people with a cancer diagnosis who want to die at home being able to do so.”