Differences in medical opinion mean some critically ill patients have just a 50 per cent chance of lifesaving emergency treatment, despite being likely to survive if they receive it.
Research by the Royal College of Physicians has found wide variations in clinicians' decisions to attempt cardiopulmonary resuscitation, which involves heart massage.
Researchers gave clinicians at an unnamed "large district general hospital" descriptions of six patients and asked whether they would attempt to resuscitate them. It also scored each fictional patient's likelihood of surviving resuscitation on three established objective scales, based on their condition.
In one case, 52 per cent of clinicians in the study said they would attempt to resuscitate the patient and 48 per cent said they would not. Yet all three objective scoring systems said the patient would be likely to survive resuscitation.
The researchers also found "concerning" evidence that clinicians had drawn opposite conclusions but still had "high levels of confidence" in their own decisions.
One patient was scored as unlikely to survive yet 38 per cent of the medics said they would attempt to resuscitate. This raised human rights concerns, the authors said, as there was a risk of inhuman or degrading treatment through the resuscitation attempt.