One in three Scottish people dies with unmet palliative care needs – what that means for assisted dying

By twelve votes, the Scottish parliament rejected the assisted dying for terminally ill adults bill on March 17. The debate that preceded it was emotionally charged and, at moments, genuinely moving. MSPs spoke of parents, partners, faith and fear. Much of it turned on the contents of the bill – safeguards, eligibility criteria and conscience clauses. The structural conditions in which terminally ill people in Scotland actually make decisions received less attention, and what attention they did receive struggled to translate into demands that any future legislation must meet. Scotland is an unequal place to live. In its most deprived communities, life expectancy has been falling since 2013; a gap of more than 13 years now separates the richest and the poorest. And the people at the bottom of that gap do not simply die younger. A 2012 study of 1.72 million Scottish patients found that having multiple long-term conditions begins ten to 15 years earlier in the poorest communities than in the wealthiest. Among the most disadvantaged of our society, the diseases that lead to terminal illness arrive sooner, in greater number and are compounded by poverty. Scotland is also, by the measures that matter most, a deeply unequal place to die. Around 6,400 terminally ill Scots spend their final months below the poverty line. One in five die in fuel poverty. The additional costs of dying (equipment, housing adaptations, heating, transport, care) amount to between £12,000 and £16,000 in the final year of life for many households at precisely the moment income collapses.

Simultaneously, new research finds that almost one in three people in Scotland die with unmet palliative care needs. Around 18,500 people a year. A separate 2024 Scottish government service mapping survey found that three NHS boards have no specialist palliative medicine doctor at all, that out-of-hours advice is unavailable in around half of Scotland’s health and social care partnerships and that over half of specialist palliative services depend on charitable rather than public funding. Following the vote, former prime minister Gordon Brown described a “moral obligation” to make urgently needed improvements to end-of-life care, warning that the “postcode lottery” means high levels of hospice and community care are available in some areas but not others. These are not footnotes to the assisted dying debate. They are its foundation. When safeguards aren’t enough The standard case for assisted dying rests on autonomy: people should be free to choose, provided they have mental capacity and are not being coerced. The Scottish bill included extensive safeguards: two independent doctors, reflection periods, requirements to discuss alternatives and enquiries into social conditions. Safeguards are designed to detect individual coercion: the controlling relative, the financial pressure applied by a family member. What they cannot detect is a different kind of pressure. That of a person who requests an assisted death, not because dying is what they want, but because the system has left them nothing else they can bear. The Equality and Human Rights Commission, which submitted evidence at every stage of the bill’s passage, put the problem precisely: “coercion or pressure is not always something applied directly by other individuals. People with disabilities may feel subtle coercion to end their lives prematurely due to attitudinal barriers as well as the lack of appropriate services and support.” It is instructive to look at Canada, which has had legalised assisted dying since 2016. Consider Sean Tagert, a Canadian man with motor neurone disease, who chose medical assistance in dying after his local health authority refused to fund the full hours of home care his doctors said he needed. The shortfall cost £200 a day, which he could not afford. He said, explicitly, that his decision was shaped by the failure of care funding. A female veteran with military service related mental health conditions reported being offered assisted dying when she asked for a wheelchair lift. A woman in Ontario died after years of failing to find housing that didn’t worsen her chronic illness, four doctors wrote to the government describing their response as “unconscionable”. In none of these cases did safeguards fail. The issue is that the safeguards were not designed to ask whether people were choosing death because every other option had been removed removed by systemic failure. In the US, Oregon has had assisted dying since 1997. Its 27-year dataset (the longest-running of any jurisdiction) provides a further, troubling signal. The proportion of patients on government insurance – a strong proxy for lower income – has grown steadily over the law’s lifetime, reaching 77% in 2024. This is nearly double the state average. Financial concerns stated as a reason for requesting assisted dying reached a record high that same year. Psychiatric evaluation, required in 27% of cases in 1998, now occurs in less than 1%. Oregon does not collect income data; it destroys case records annually. Patterns of inequality are difficult to find when nobody is looking for them. I do see Scotland returning to this question (the bill failed by only twelve votes). When it does, the inequality argument must do more than determine how people vote. It must shape what is proposed. Any future bill deserves scrutiny not just of its safeguards, but of the conditions those safeguards operate in: whether palliative care is genuinely available, whether dying people are financially supported, and whether the data exists to know, in real time, whether structural disadvantage is shaping who requests an assisted death and why. A choice made because there is no other bearable option is not a free choice. In Scotland today, for thousands of people at the end of their lives, that is precisely the situation. The assisted dying debate has the wrong question at its centre. The right one is: what kind of dying does Scotland currently provide, and for whom?

Philip Broadbent receives funding from the Wellcome Trust Multimorbidity Doctoral Training Programme 223499/Z/21/Z