Atheist or agnostic doctors are almost twice as willing to take decisions that they think will hasten the end of a very sick patient's life as doctors who are deeply religious, suggests research published online in the Journal of Medical Ethics.
And doctors with a strong faith are less likely to discuss this type of treatment with the patient concerned, the research shows.
The findings are based on a postal survey of more than 8500 UK doctors, spanning a wide range of specialties, which was designed to see what influence religious belief - or lack of it - had on end of life care.
The specialties included those in which end of life decisions would be particularly likely to arise, such as neurology, elderly care, palliative care, intensive care and hospital specialties, and general practice.
The doctors were asked about the care of their last patient who died, if relevant - including whether they had provided continuous deep sedation until death and whether they had discussed decisions judged likely to shorten life with the patient - their own religious beliefs, ethnicity, and their views on assisted dying/euthanasia.
Nearly 4000 doctors responded (42% of the total surveyed), and almost 3000 reported on the care of a patient who had died.
Specialists in the care of the elderly were somewhat more likely to be Hindu or Muslim, while palliative care doctors were somewhat more likely than other doctors to be Christian, white, and agree that they were "religious."
But, overall, white doctors, who comprised the largest ethnic group among the respondents, were the least likely to report strong religious beliefs.
Ethnicity was largely unrelated to rates of reporting ethically controversial decisions, although it was related to support for assisted dying/euthanasia legislation.
Specialty was strongly related to whether a doctor reported having taken decisions, expected or partly intended to, end life. Doctors in hospital specialties were almost 10 times as likely to report this as palliative care specialists.
But irrespective of specialty, doctors who described themselves as "extremely" or "very non-religious" were almost twice as likely to report having taken these kinds of decisions as those with a religious belief.
The most religious doctors were significantly less likely to have discussed end of life care decisions with their patients than other doctors.
These attitudes were reflected in support for assisted dying/euthanasia legislation, with palliative care specialists and those with a strong faith more strongly opposed to it. Asian and white doctors were less opposed to such legislation than doctors from other ethnic groups.
The author concludes that the relationship between doctors' values and their clinical decision making needs to be acknowledged much more than it is at present.