Terminally Ill Adults (End of Life) Bill

Baroness Garden of Frognal Excerpts

Friday 30th January 2026 Lords Chamber

[Baroness Finlay of Llandaff (CB)]

My Lords, this group concerns terminal illness. It is incredibly important. It is the title of the Bill, so it is extremely important that we have proper definitions of terminal illness in the Bill that are beyond doubt. The proposal to make eligibility dependent on a clinician’s judgment that a person is expected to die within six months might appear to be a safeguard, but evidence from palliative care, primary care and prognostic research shows that predictions are highly unreliable and largely guesswork. Some national celebrities who clamoured for eligibility live for years, while others campaigning for this Bill for themselves are clearly ineligible. Do doctors need to be 95% confident or 50% confident to rule that the patient will die within six months to be eligible? No validated clinical tools or tests exist to make such determinations with sufficient reliability for life-ending decisions. This lack of reliable knowledge leaves clinicians to face pressure to make prognoses they cannot substantiate with concurrent legal risks and ethical dilemmas.

Medical science recognises broad trajectories of decline. Those dying of cancer often experience a relatively long stable period followed by rapid deterioration over weeks or months. People with advanced organ failure, such as heart or lung disease, tend to decline in an unpredictable, fluctuating pattern marked by crises and partial recoveries. Others, particularly those with multiple conditions, neurological illness or frailty, experience a slow, progressive dwindling over years with an average of 20 or more months from identification of early palliative care needs to death and, during that time, quite marked fluctuations.

While these patterns help clinicians anticipate needs for care, they do not forecast how long an individual has left to live. A six-month prognosis is not a clinical truth. Medical prognostication for terminal illnesses is based on averages. However, in the individual, marked variables influence survival, including treatment response, infections and physiological resilience. A time-based legal threshold would turn population-level averages into high-stakes individual errors, approving some who are not within six months and denying others who are. An assisted death is irreversible, amplifying the consequences of prognostic errors and litigation. A systematic review of 42 studies on clinicians’ predictions for terminal cancer patients found very low accuracy rates ranging from 23% upwards. In a multicentre study of more than 2,000 palliative care patients with advanced cancer, physicians’ clinical predictions were accurate in only 35% of cases. Tools such as the surprise question—Would you be surprised if this patient is alive in a certain number of months or weeks?—have been tried as a way to predict those likely to die. It performs poorly at six to 12-month predictions, especially for people who have multiple conditions.

The six-month rule is unsafe and unworkable. Assisted dying legislation requires criteria that are objective, transparent and as evidence-based as possible. The definition must exclude conditions that can be meaningfully stabilised or controlled for years, as in Amendment 104 in the name of my noble friend Lady Grey-Thompson. Manageable conditions must not be treated as terminal if somebody during an episode of depression or other life trauma, such as bereavement or a relationship break-up, decides to stop whatever treatment they have when they are feeling suicidal, such as the diabetic who decides to stop their insulin knowing that if they do, they will die within weeks.

The arbitrary time of six months is unverifiable. You have a less than 50% chance of being right, and the index of suspicion, as written in the Bill, is far too low. A review paper by Scott Murray and Simon Etkind, published in the British Medical Journal in March 2025, described how inaccurate prognostic tools are—as well as the importance of illness trajectories, which can

“provide broad timeframes and patterns of likely decline”

to help doctors understand disease progress.

Patients with frailty have been described as being “on a tightrope”: some will stay on for years, but others may suddenly be blown off by an infection or a fall, and die rapidly. The Bill says nothing about pain, suffering or potential suffering. The only disease-related issue in it is prognosis, yet accuracy at six months below 50% is a pure guess that someone has six months to live. To tackle these problems, Amendments 71, 77 and 90 offer options to look at the progress of disease to increase the chance of being a little more accurate at six months, which is the point at which you could trigger the process of applying for an assisted death.

Amendment 79 would help clarify whether the disease is progressing because some potential disease-modifying treatments have been ignored—or, perhaps, the patient does not even know about such treatments. The progress of the illness will give a far better indication of whether you are rapidly deteriorating or are, in fact, remarkably stable.

My Amendment 444A would allow you to be monitored while you have other support, having triggered a request for an assisted death. This would allow specialist palliative care to be available to you if you wish to try out various suggestions. Then, as the disease is progressing more obviously towards death and there is a greater sense of predictability, the person would be eligible for sign-off by the panel and given the lethal drugs. The best evidence on predictability would be the point at which the palliative prognostic index score indicates that the person has a more than 50% chance of dying within six weeks.

However, Amendment 444A offers a compromise, because I am sure that many people will feel that six weeks is too short a timeframe. It suggests that the patient cannot have lethal drugs until they really are seen to be terminally ill and the clinicians are more confident in predicting that death is likely to occur within three months. The patient who is rapidly declining would not need to wait between first assessment and having an assisted death because they would be seen to be declining rapidly, but the patient who is completely stable would not risk having the lethal drugs until there was evidence that they were actually in the last phase of their terminal illness and a wrongful death could be avoided.

Some jurisdictions include frailty in their definition of “terminal illness”, but others do not. Evidence on this came up in Committee in the other place: a 2025 paper in Age and Ageing titled, “Why We Need to Consider Frailty in the Assisted Dying Debate”. With this extensive uncertainty around frailty and other multimorbid conditions, does the noble and learned Lord, Lord Falconer, include frailty in his definition of “terminal illness” or not? Also, does he accept that disease progress is critical to avoiding wrongful deaths? I beg to move.

[The Deputy Chairman of Committees (Baroness Garden of Frognal) (LD)]

My Lords, the noble Baroness, Lady Campbell of Surbiton, is taking part remotely. I invite the noble Baroness to speak.

[Baroness Campbell of Surbiton (CB) [V]]

My Lords, before I speak to the amendments in this group, I say that I too have not received the letter sent by the noble and learned Lord, Lord Falconer, regarding amendments he proposes to return with, so I am somewhat in the dark today, should the content relate to remarks I am about to make now. Please would he check the distribution list, as such information is crucial to our deliberations, especially when one relies so heavily on remote participation for inclusion, as I do?