The year he retired from full-time practice, in 2014, Henry Marsh published Do No Harm, a likable and arrestingly honest memoir about his 40-year career as a brain surgeon. Four years earlier, while still a consultant at St George’s hospital in south London, he had reviewed the debut work of another doctor-writer, Siddhartha Mukherjee, an Indian-American oncologist who, in The Emperor Of All Maladies, explored in incredible depth the history of cancer and its treatment. (Marsh called that book “exhaustive, and exhausting”. It went on to win the Guardian First Book award and a Pulitzer prize.)

Mukherjee is now 47 and lives in New York; Marsh, 67, lives in Oxford. To different extents both of these doctors still practise in their respective fields – Mukherjee at Columbia University’s cancer centre, Marsh as a visiting doctor at various hospitals around the world, including in Kathmandu in Nepal. Both men have continued to write: Marsh a second volume of autobiography, called Admissions, published this year, and Mukherjee a study of genetics called The Gene: An Intimate History, published last year. When they sat down to talk to each other over Skype one Saturday afternoon in November, they began with a subject on which their two lifelong disciplines overlap: the treatment of brain cancer. Tom Lamont

Henry Marsh You look as though you live in some sort of Mies van der Rohe building – a Philip Johnson house, or something.

Siddhartha Mukherjee Well, this is our house in the country. It has no furniture in it.

HM Modernism! Right. Well, I reviewed your cancer book a couple of years ago. I thought it was wonderful, really very interesting indeed.

SM I’ve now read both of your books.

HM What I’d like to start talking about is something that bothers me greatly. I still do one day a week in my hospital in London, when I’m not working abroad. I was sitting at a tumour board meeting – a difficult case, an elderly man – and my oncology colleagues said, “We’ll treat him.” I asked, “Isn’t there a case for doing nothing?” They said, “No, we’re told if they’ve got a prognosis of more than six months, you’ve got to treat them. We can’t be ageist.” I said, “What if he was 100 years old?” My question for you, Sid, as an oncologist, is: when do you stop?

SM Obviously it’s a question we’re grappling with, not only as physicians but as a country, a society. I think, first of all, patients themselves know when to stop. One good guide is to ask them. It’s an old idea in medicine, that the first question you ask, when you’re trying to diagnose something, is: “What do you think it is?” And he or she might not be correct, but they often have an instinct. This is not to move the onus of responsibility away from doctors. But usually when we ask people, “What do you really want to do?” and you paint an honest picture of what it’s likely to be like, I think most people make decisions that are within the bounds of reason.

HM But do you think most oncologists do that? Do they paint that honest picture?

SM I think there are pressures…

HM Because I do sympathise. I’ve spent most of my work conveying risk and probability to people with brain tumours.

SM I think the crux is the word “honest”.

HM Exactly.

SM Here’s something that I took from your first book, and that I believe in: the relationship between doctors and patients will always be a relationship of unequal power.

HM Yup. Agreed.

SM And it’s not as if you can equalise it. The patient is the one with the illness, they’re the one who has come to you for advice and counselling. So none of this is to say: move the responsibility of decision-making on to the patient. He or she doesn’t know. He or she is asking you for advice. But I think it’s important, particularly in this day and age, to paint an honest picture for them of what is most likely to happen.

HM It’s all about probabilities, isn’t it? To sit down with a head-injury patient and say to the family, “Well, actually, your nearest and dearest is probably better off dead” – that’s terribly difficult. One always is worried: I might be wrong. And as I got older, I think I got better at accepting, well, I might be wrong. But it’s very unlikely. It’s wrong to treat everybody with a head injury for the sake of maybe one good result.

It’s very difficult with younger doctors. They want to treat. They want to operate. And it’s so frightening, it’s so difficult, to deprive people of hope. It’s a long, painful conversation – often in the middle of the night.

SM In surgery, you are making a decision that will either commit or not commit: open a human being, or don’t. In the case of cancer, where the decisions can be renegotiated at every stage, it’s different. You’re not committing yourself to a decision or decisions that can’t be reversed.

HM I think all of this leads on to a very deep ethical problem, in terms of training and learning. Because, as a young surgeon, if you don’t take on the difficult cases, how is one ever going to get any better? You have a duty to the patient in front of you, but also a duty to future patients.

SM Similarly, when we’re asking patients to enrol in clinical trials, where the chances of it working are anything from 5% to 50%, we’re actually asking them to be involved in an experiment. The word “experiment” makes it sound much colder than “clinical trial”. Culturally, we’ve got used to the phrase “clinical trial”, but these are medical experiments, and we know medical experiments can go horribly wrong and can yield very little in terms of working value.

HM You told a story that I was very struck by, in your cancer book, about bone-marrow treatment for breast cancer. How it was pushed by a big publicity campaign, before proper trials, and if I remember correctly, the trials showed it was a bad idea.

SM And in fact we’re still living through the legacy of that. It created such an aura of hope, and optimism, and when that was destroyed, so much was destroyed with it. And one of the problems there was that people thought they knew the answer before the answer was there. And that’s a terrible formula.

HM How old were you when you left India for the States?

SM I was 18. I went to college at Stanford.

HM Do you think that’s given you a certain outsider’s detachment, in some ways?

SM I came to the States not only to be technically trained: I came because I was curious about another world.

HM I would say that, because I went late into medicine [in his mid-20s], I’ve never entirely identified with the medical profession. I have a certain sense of being an outsider. I think it’s helped me write in the way that I have done. I suspect the same applies to you.

SM The idea of being curious about a different world, I think, fundamentally describes my writing practice.

HM I became a formal writer at a much later age than you did. But I’ve been writing all my life, and I’ve always kept a diary, which became a sort of therapy for me, actually. I knew that I lived an interesting life, and that I was seeing extraordinary things, and that I didn’t want to feel as if this was sand slipping through my fingers. I don’t know how your writing started, Sid?

SM You know, I write to think. That’s my method of trying to figure out the world. Why are we here today? What will happen five, 10, 50 years from now? The only way to do that, me, has been to write.

HM As you know, at the end of my second book I make a passionate plea for doctor-assisted suicide. Of course, there are several American states now that have agreed to it. I think the American physicians’ association has agreed to it. What is your own view, as an oncologist?

SM There are some legal and cultural barriers which are enormous. In some ways, my fear about the way it’s moved in the US is that there were firebrands such as Jack Kevorkian [an American proponent and practitioner of doctor-assisted euthanasia who was convicted of second-degree murder in 1999, and who died in 2011].

HM He was pretty sick, I think.

SM That’s right. So, in restarting the conversation, “What does it mean to die, today, in America, in a way that’s dignified and painless?” – that conversation’s been so contaminated by what happened all those years ago that it’s very difficult, politically, to come back to it.

HM It’s rather like in Germany, where you can’t discuss it at all, because of the Nazis’ euthanasia programme in the 1930s.

SM So I think the way forward is to try to start from relatively basic principles. Which is that, in the next decade or so, science will invent medicines that will progressively make pain a profoundly handleable symptom. We already have powerful medicines, and we will have more powerful ones. So the idea of a painful death – if we could take the pain part of the equation away, then you are left with fundamental questions such as: what is dignified?

HM My father died at 96, 10 years too late. He had Alzheimer’s, and the man he was when he was demented was not the man he was before. It was a relief when he died. But my mother, she died at the age of 82, from suddenly progressive metastatic cancer, and it was a wonderful death. I mean, death’s always sad, but she died in full possession of her faculties. A bit drawn out, but she was at home. I felt we were very lucky.

SM Some of the toughest decisions, when my own father died, were about approaching the conversation about withdrawal of care with my mother. I dreaded that conversation. I have that conversation all the time with my patients. And yet the amount of dread that brought on me, when it was my turn to have that conversation as a son and not a doctor, was astonishing. There were things I knew as a doctor that I didn’t know as a son. There were moments in which I was making decisions as a doctor when I should have been making decisions as a son. And moments when I was making decisions as a son when I should have been making decisions as a doctor. As you will know, Henry, doctors make terrible patients.

HM Yup.

SM It’s because they – we – cannot do unto ourselves what we can do unto others. That’s been a historical failing.

HM I think it’s all about the way we have to develop this split personality, and be emotionally detached from patients; it makes being a “medical relative” incredibly hard. I’m a great admirer of Atul Gawande’s books, especially Being Mortal. But I was very puzzled at the end of that book when he described his father’s death. His father had astrocytoma [a type of cancer of the brain]. There’s no effective treatment for it. And yet they go down the route of surgery, radiation, chemotherapy.

SM I’m also a big admirer of Atul. We’ve become friends and we’ve had multiple conversations around these topics. I think, again, that one of the failings of oncologists is that they don’t present to patients a landscape of what might happen.

HM Rather than saying, “If you’re lucky, it could all be great.” I read a book recently called The Violet Hour by Katie Roiphe, about famous writers’ last few months of life, and one of the subjects was Susan Sontag. [Sontag died in 2004, from complications of myelodysplastic syndrome, which had evolved into a form of leukaemia.] As far as I could tell, she kind of tortured herself to death, offered more and more heavy experimental chemotherapy. It was presented [in Roiphe’s book] as a sort of brave struggle, but I thought it was groteseque, slightly crazy almost.

SM I’ll tell you what I do in my own practice. I tell stories, actual stories, about actual cases. This allows people to understand where they might really fit. They don’t have to think of themselves in terms of numbers.

HM Oh, numbers are hopeless. It’s a bit different in America from England, because in England, being a socialised healthcare system, basically you’re stuck with the doctor you go to. In America, patients are more like consumers and shop around a lot more.

SM I can’t stand that, the patient-as-consumer. It makes stupid a relationship that is extraordinarily complicated and can be extraordinarily hard to negotiate. And it demeans the relationship you have with the patient.

HM It sounds condescending, but the fact of the matter is that patients do not know what’s best for them.

SM I think the cases that are most successfully treated – and I don’t mean just survival…

HM Death can be a good outcome, sometimes, sadly.

SM Quite. So those cases where the families or patient has the most closure, those cases are typically ones in which there’s been a longitudinal relationship built with one doctor. And this goes back to the shamanic quality of medicine. We inherit that culturally, we inherit that anthropologically, we inherit that as human beings, and that’s not going to change.

HM I’ve always felt the dissection we do as medical students is of more value in terms of an initiation rite, rather than in terms of teaching us anatomy. I was amazed to read recently that, at one medical school in America, they were anxious to make their medical students more empathic, by making sure the students met the living families of the dead bodies they were going to dissect. It struck me as a completely crazy idea, in that it was undoing the very value of dissection. Which is to become detached. The problem is, you mustn’t become too detached.

SM I was struck, recently, by an article written by Zeke Emanuel, in which he wrote about his decision to phase out medical care past a certain age. The article inspired a way of thinking. “Look, the way I want to run my medical care is as follows. At 80 years old, I want to discontinue the following procedures on myself. At 85, I want to discontinue yet another few. At 90, I want to discontinue yet another few. And only allow a few remaining ones.” It is a kind of phase-out, of things we will allow and won’t allow. And what’s thoughtful about the phase-out, to me, is that it takes what seems like an all-or-nothing decision and allows you basically to say: “My life is closing down. These are my options. I’m going to turn things off one by one.”

HM But this guy’s a doctor, so he’s a very informed customer… Bad word, customer! But you know what I mean.

SM The idea is to have some reason about treatment when you’re 80.

HM Yes: to think a bit about the future as you get older. Under the NHS, everybody over the age of 60 has their blood pressure taken regularly by their GP, and various other tests. I’d like to see a system where everyone over the age of 65 or 70, the GP gets them in and says, “Look, can we discuss your future healthcare? What do you want or not want?” But when I’m dying myself, how will I feel? I’m sure I’ll be as frightened and anxious as the next person.

SM Absolutely. Yes.

• Commenting on this piece? If you would like your comment to be considered for inclusion on Weekend magazine’s letters page in print, please email weekend@theguardian.com, including your name and address (not for publication).