Palliative care: the conversation you're having too late

Clinical Rounds, the GPAtlas podcast. Palliative Care. 19 minutes. Published 31 August 2026. Free to listen.

Clinical source: our own Ocean article "Palliative care - general issues", last reviewed 6 August 2026.

Transcript

Sarah: Hello, and welcome to GPAtlas Clinical Rounds. I'm Sarah, and this is the podcast where we take one thing you'll genuinely see this week and work out what to do with it.

Ben: And I'm Ben. Today is palliative care, and specifically the general running of it, which is almost entirely our job rather than anybody else's.

Sarah: So. Mr Doyle is seventy-nine. Lung cancer, diagnosed eighteen months ago, and he's had his last cycle of chemotherapy. He's still coming to the surgery, he's still driving, and he looks reasonably well. His wife rings on a Tuesday to ask whether he should have his flu jab. That's the whole consultation. And the reason I'm starting there is that this is the appointment where palliative care either starts or doesn't.

Ben: Three things today. When to start, which is the bit we get wrong. What to actually put in the house before you need it. And what to hand over, so that a bad night doesn't automatically become an ambulance.

Sarah: So, when. Our article has a line about this that I found uncomfortable reading, because it names the failure precisely. Late identification, rather than lack of treatment, is the commonest failing in community palliative care. Not that we treat badly. That we start late.

Ben: And the tool for that is the surprise question, which most people know: would you be surprised if this person died in the next twelve months? What our article does that's more useful is spell out what happens when the answer is no. It isn't a register entry. It's five things. Add them to the practice palliative or supportive care register, and then act on it: start advance care planning, prescribe anticipatory medicines, complete an SR1 for benefits, and hand over to out-of-hours.

Sarah: And I want to sit on the SR1 for a second, because it's the one people don't associate with palliative care at all. It replaced the DS1500 in twenty twenty-five, and the trigger is exactly the same surprise question: you complete it when you would not be surprised if the person died within twelve months. It fast-tracks Personal Independence Payment, Attendance Allowance and the related benefits. So for Mr Doyle, sitting there asking about a flu jab, that form is probably already appropriate. And nobody has done it, because he looks well.

Ben: The failing is late identification. Not undertreatment.

Sarah: Right. Anticipatory prescribing. The just-in-case box. Four medicines, and it's worth knowing them as a set with their indications, because you will be asked to authorise this at four in the afternoon with a queue outside.

Ben: Four subcutaneous as-required medicines, in advance, for the home. An opioid, morphine, or oxycodone if there's renal impairment, and that covers pain and breathlessness. Midazolam, for agitation, breathlessness or seizures. An antiemetic or antipsychotic, so haloperidol or levomepromazine. And an antisecretory, hyoscine butylbromide or glycopyrronium.

Sarah: And the detail that decides whether any of that works is not the prescription. It's the logistics. Our article says to make sure the medicines and an administration chart are in the home before they're needed, if deterioration is expected. Before. Because a prescription signed on Friday afternoon that's still sitting at a pharmacy on Saturday night is not anticipatory prescribing, it's paperwork.

Ben: Two numbers to carry. The breakthrough dose for an opioid is one-sixth of the total daily dose, and you always prescribe one. And the syringe pump trigger: consider a pump if more than two to three as-required doses are needed in twenty-four hours. That's the number that tells you the oral, intermittent approach has stopped working.

Sarah: One-sixth for breakthrough. Two to three doses in a day means think pump.

Ben: Now the prescribing error that our article calls out, and it's the commonest one. Morphine in renal impairment. If the eGFR is below thirty, you reduce the dose, lengthen the interval, and switch, to oxycodone in mild to moderate impairment, or alfentanil or fentanyl in severe. And you seek specialist palliative care advice. Morphine metabolites accumulate, and the patient gets agitated and twitchy, and it looks exactly like the disease progressing.

Sarah: Three smaller ones that come up constantly. Co-prescribe a stimulant laxative with every regular opioid, and warn about transient nausea and drowsiness, because if you don't warn them they stop the opioid. Patches are not for rapid titration or for opioid-naive acute pain, and heat and fever raise fentanyl absorption, which is an MHRA safety concern. And metoclopramide is licensed for short-term use, up to five days, because of extrapyramidal effects, and it's avoided in complete bowel obstruction and in Parkinson's.

Ben: And on the antisecretory, there's a choice that matters. Hyoscine butylbromide, so Buscopan, or glycopyrronium, is first-line, and both are preferred over hyoscine hydrobromide, which crosses the blood-brain barrier and can cause sedation and delirium. Same family, very different effect on whether your patient is present in their last days.

Sarah: And on the rattle itself, our article says something I've started repeating to families. You reposition, and you reassure the family, because it rarely distresses the dying person. The noise is almost entirely a problem for the people in the room. Saying that out loud, gently, does more than the injection does.

Ben: Below an eGFR of thirty, morphine is the wrong opioid.

Sarah: The handover, which is the part that decides whether all the rest of it survives contact with a Saturday night. Advance care planning, anticipatory medicines, and an out-of-hours handover, so that a crisis does not default to hospital. That phrase, default to hospital, is exactly what happens when nobody has written anything down.

Ben: Same for the resuscitation decision. Complete and communicate a valid plan, DNACPR or ReSPECT, and share it with out-of-hours and the ambulance service. A form in the notes that the crew never sees does nothing at three in the morning. And before you make any best-interests decision, check whether there's already a valid advance decision or a health and welfare lasting power of attorney, because if there is, that's the decision, not yours.

Sarah: And the carer, who is the patient you didn't book. Identify and support them, arrange respite, equipment and night sitting, and signpost to hospice and charity services. Our article puts sudden carer breakdown in its referral list as an urgent respite or community support need in its own right. Because when the carer goes, the whole plan to die at home goes with them, that day.

Ben: If out-of-hours cannot see it, it does not exist.

Sarah: Two emergencies to name, because they turn up in exactly this group and they are not palliative situations, they are acute ones. Metastatic spinal cord compression: cancer, plus new back pain with neurological signs, is an emergency, sixteen milligrams of dexamethasone and same-day acute oncology or the MSCC coordinator. And neutropenic sepsis: fever, or just feeling unwell, within six weeks of chemotherapy is an immediate emergency admission, and you do not wait for bloods.

Ben: And the conversation that underpins all of it. Our article describes explaining in plain language that anticipatory medicines treat distressing symptoms, and that correctly dosed opioids relieve suffering rather than shorten life. That's the double effect principle, and families ask about it directly, usually as, will the morphine finish him off. Having a clear, unembarrassed answer ready is part of the prescribing.

Sarah: Cancer, new back pain, neurology. That is today, not palliative.

Sarah: Let me put a second patient in, because the first one was the easy version. Mrs Iqbal is sixty-eight, ovarian cancer, and she's deteriorated over about ten days. She's in bed now, she's stopped eating much, and she's drowsy for most of the day. Her daughter rings on a Thursday morning and says, I think something has changed.

Ben: And recognising that is a clinical skill we don't teach well, partly because it's mostly recognising a pattern rather than measuring anything. What our article does give you is the management once you've recognised it. Use the anticipatory medicines and a syringe pump if the oral route fails, or in the last days of life. Start at the lowest effective dose, titrate to response and side effects, and reassess at least daily in the last days.

Sarah: At least daily. That's the standard, and it's worth saying because it's the thing that quietly doesn't happen. A visit on Thursday, a plan, and then the weekend. Whereas what that phase actually needs is somebody looking again every day, because doses that were right on Thursday are often wrong by Saturday.

Ben: And it changes what the family is being asked to do. They're now giving subcutaneous medication in a bedroom, watching for effect, and deciding when to call. Which is why the administration chart being physically in the house matters as much as the drugs, and why the out-of-hours handover has to say what the plan is, not just that there is one.

Sarah: At least daily in the last days. That is the standard.

Sarah: And the conversation the daughter is actually having, underneath the clinical one. She asks whether the morphine is making her mother sleepy, and what she means is, am I helping to end this. Our article's phrasing is the one I use: correctly dosed opioids relieve suffering rather than shorten life. That's the double effect principle, and it deserves a straight answer rather than a reassuring noise.

Ben: And it's worth being equally straight about what the medicines are for. Anticipatory medicines treat distressing symptoms. They aren't a countdown, and having them in the house doesn't commit anybody to anything. Families sometimes refuse the box because they think accepting it means giving up, so saying that explicitly, when you prescribe, saves a bad night later.

Sarah: Two practical things that make the difference between a plan and a working plan. Equipment, so a bed, a mattress, a commode, which takes days to arrange and therefore has to be started before it's obviously needed. And night sitting, which is the single thing most likely to keep somebody at home, because it's the nights that break carers, not the days.

Ben: The box does not commit anyone to anything. Say so when you prescribe it.

Sarah: Let me say something about the register itself, because it's the bit of practice machinery that makes all of this work or not work. Our article puts adding someone to the practice palliative or supportive care register as the first action after the surprise question. And a register is only useful if somebody looks at it. If it's a list that gets updated when someone dies, it's an archive, not a register.

Ben: And the reason it matters is that everything else on that list of five hangs off it. The advance care planning conversation, the anticipatory medicines, the SR1, the out-of-hours handover. Those aren't four separate tasks somebody has to remember. They're what the register entry is for.

Sarah: And hospice, which is the other thing people leave late for the same reason they leave everything late, because referring feels like conceding. Our article puts hospice and charity services alongside respite, equipment and night sitting as the support package, not as the final step. Marie Curie is named in it. Most hospices take referrals far earlier than GPs assume, and for symptom control rather than for dying.

Ben: A register nobody reads is an archive.

Sarah: And the last thing, which is about us rather than the patient. This is one of the few areas of general practice where you are still genuinely the main clinician. Not the coordinator of other people's plans, the person actually deciding. Mr Doyle's flu jab appointment, the one we opened with, is worth the extra four minutes precisely because nobody else is going to have that conversation with him.

Ben: And the failing our article names is the one you can fix on a Tuesday with no extra resource at all. Late identification, rather than lack of treatment. You already know how to prescribe the four drugs. The question is only whether you started six months earlier than you would have.

Sarah: You already know the drugs. The question is when you started.

Sarah: And one practical thing I'd add about the anticipatory box, which is what happens if it's never used. Quite often it isn't. Somebody dies peacefully, the drugs sit in the house unopened, and there's a version of that which looks like waste.

Sarah: It isn't. The box is insurance, and the cost of not having it is a crisis at two in the morning with no drugs in the house and an out-of-hours clinician who has never met the patient. Our article's framing is to have them there before they're needed if deterioration is expected, and unused is the best outcome, not a failed prediction.

Ben: Which is worth saying to the family too, when you prescribe them, because they can read the arrival of four injectable drugs as a statement about how long he has. Telling them plainly that most boxes are never opened, and that having it changes nothing about the timeline, takes the sting out of it.

Sarah: Unused is the best outcome. Not a failed prediction.

Ben: Exam corner. A seventy-two year old man with metastatic prostate cancer is at home on modified-release morphine, sixty milligrams twice daily. His pain is mostly controlled but he gets breakthrough pain twice a day. What breakthrough dose do you prescribe? A, ten milligrams. B, twenty milligrams. C, thirty milligrams. Or D, sixty milligrams.

Ben: It's B, twenty milligrams. His total daily dose is a hundred and twenty milligrams, and the breakthrough dose is one-sixth of the total daily dose, which is twenty. A is a sixth of a single dose rather than the daily total, which is the commonest error and leaves him under-treated. C and D are too high and will sedate him. And note he's having breakthrough twice a day, which is under the two to three doses in twenty-four hours that would make you think about a syringe pump, so for now you treat the breakthrough and review.

Sarah: So. Three things for Monday. One. Ask yourself the surprise question about somebody who looks well, because late identification is the failing, not undertreatment. Two. If the answer is no, that's five actions, not a register entry: planning, anticipatory medicines, the SR1, and the out-of-hours handover. And three. Check the eGFR before you write morphine. Below thirty, it's the wrong drug.

Sarah: One thing to reflect on, if you're logging this. Think of one patient on your list who would not surprise you. Have they got a handover, and have they got the box in the house? The full transcript and the references are on the episode page. Ben and I are synthetic voices. The medicine isn't. See you next week.

More from Clinical Rounds

Every episode of Clinical Rounds, free to listen with a full transcript.