Once, when I was still only a medical student, I was told by an attending physician to go and tell the relatives of a patient, who had cancer of the lung, that he was dying. Unhappily for me, the man’s nearest relative, his wife, was almost totally deaf, and so before long I found myself bellowing the news at the top of my voice for the whole hospital to hear. I remember thinking that this was not the way things should be done.

But what is the way things should be done? An essay in the New England Journal of Medicine for May 3, titled “Freedom from the Tyranny of Choice – Teaching the End-of-Life Conversation,” points out that doctors are still largely at a loss about what to do when a patient’s death is close but he could be kept alive a little longer with active  and intrusive treatment. Should they present the patient, or the patient’s relatives, with a kind of multiple choice list and let him or them decide among the various options? “Do you want to be resuscitated?” is an embarrassing question to have to ask anyone; doctors often get round it by asking people auxiliary questions such as “Do you want further lab tests?” or “If we turn off the insulin drip, is it all right if we don’t check the blood sugar?”

The essay suggests that the conditions of modern medical practice are not conducive to proper end-of-life conversations, as it delicately calls them. The physician who knows the patient best, and whom the patient trusts most, is often — usually — not available; thanks to increasingly stringent regulations concerning their working hours, residents rarely get to know patients intimately enough to engage in such conversations. And so, in the end, everything is done in an atmosphere of hurry and embarrassment, if it is done at all. In one survey of 2500 patients with metastatic cancer, only 20 percent had any clear plan or instructions as to what to do when the end was near.

The authors of the essay write: “Although we learn guidelines for using antibiotics and managing acute coronary syndrome, in our experience, we do not learn similar guidelines for end-of-life conversations.” I confess that at this point I began to smile, if a little wanly. I knew that a checklist was coming, and so it was: the “step-by-step SPIKES approach.” SPIKES stands for “setting up, perception, invitation, knowledge, emotions, strategy and summary.” The authors add: “They are not widely used.” I am not surprised.