Stay informed with free updates
Simply sign up to the Science myFT Digest — delivered directly to your inbox.
The writer is a science commentator
On Friday the UK parliament will vote on whether terminally ill individuals with mental capacity can ask for medical help to hasten death. If the assisted dying bill passes, the country will move closer towards sanctioning state-assisted suicide, bringing it in line with countries such as Switzerland, Belgium and Canada.
There are compelling arguments on both sides of this emotive issue and the result feels impossible to call. While public opinion broadly supports change, parliamentarians seem more divided.
But one aspect has been overlooked: the science of predicting survival. As a safeguard, the bill restricts the right to assisted dying to those with less than six months to live. While ballpark estimates of survival can be calculated using patient groups, forecasts for individuals are harder to pin down.
“My research demonstrates that there is no reliable way to identify patients with less than six, or twelve, months to live . . . at least, no method that would be reliable enough to act as any sort of ‘safeguard’ for the proposed assisted dying legislation,” Paddy Stone, emeritus professor and former head of the Marie Curie palliative care research department at University College London, told me this week.
According to Nicola White, a senior research fellow in the same unit, health professionals are no more able to offer accurate timelines of survival today than they were 30 years ago, even with the aid of additional markers like blood tests. While forecasting survival for cancer patients is fraught enough, it is even tougher with heart failure and neurodegenerative disease. These important challenges — accurately predicting an individual’s survival and deciding what counts as a terminal illness — have been overshadowed by other issues, such as the role that judges will play.
Predicting survival is like weather forecasting: the closer a patient comes to the end of life, especially the last 24 hours, the easier it becomes to estimate prognosis. Forecasting the last seven days of life is harder than the final 24 hours; further out than that, things become shakier still.
To test accuracy, studies sometimes ask doctors to say whether a patient nearing the end of life has “days”, “weeks”, or “months” left. One 2023 paper covering about 98,000 patients showed that clinicians were 74 per cent accurate in judging who would live for less than 14 days and 83 per cent accurate on who would survive more than a year. But that fell to 32 per cent when estimating those in the middle, likely to live “weeks” or “months”.
“All the studies from this country and others show that estimating [whether a patient has] six months left to live is extremely difficult and not that accurate,” says Irene Higginson, professor of palliative care and policy at King’s College London and scientific director of the charity Cicely Saunders International. “The science isn’t that well developed and I’m not sure it could be, because individuals vary so much.” Higginson declined to offer her opinion on the bill.
Many palliative care professionals fear that assisted dying will suck resources away from end-of-life care. Last month, the Association for Palliative Medicine voiced opposition because of concerns about protections for the vulnerable; inadequate provision of end-of-life care services across the UK; and the impact on trust between doctor and patient. Higginson points out that palliative care already includes the right for patients to refuse medical treatment.
Opponents wave at other countries as proof of a slippery slope: Belgium and the Netherlands now permit euthanasia for under-18s; in Canada, the lonely and the homeless have asked to die. One Toronto university psychiatrist described the country’s “medical assistance in dying” legislation as a “bait and switch”, with a well-intentioned law “metastasising” into something malign.
Those who champion assisted dying do have good intentions. They cite patient autonomy and human rights. They argue that a good death should not be restricted to those able to afford a trip to Dignitas.
The point of this column, though, is not to argue for or against the assisted dying bill — but to ask whether the science underpinning one safeguard delivers what is asked of it. The answer? Possibly not. Ultimately, parliamentarians might regard the scientific uncertainty in survival prediction, among other difficulties, as a trifling consideration when set against the opportunity to deliver liberalising reform and patient choice.
And that’s the point, really: this profoundly important vote should be an informed one.