Content area
Full text
Introduction
Accurately recognizing that a person may be dying and in the last weeks or days of life is central to improving peoples’ experience of care. It enables families, medical teams and health-care providers to plan and provide the best care possible. However, physicians’ predictions of dying are frequently inaccurate and overoptimistic1. The 2019 United Kingdom National Audit of Care at the End of Life in hospitals found the recognition of the dying was challenging. When dying was recognized, 20% of people died within 8 h; the median time to death was 36 h; and importantly 50% of patients lacked the capacity to be directly involved in any decision-making2.
Predicting when a patient with advanced cancer is likely to die is a challenge and currently no diagnostic test is available. Globally, there were 19·3 million new cancer cases and almost 10 million cancer deaths in 2020; lung cancer had the highest mortality, responsible for 1·8 million deaths3. Accurate prognostic information at the end of life is essential to co-ordinate and manage care in response to need, whilst avoiding burdensome and unnecessary interventions. Several validated prognostic tools aim to predict the survival of patients with advanced cancer4. A recent comparison of five validated prognostic tools showed the best model, PiPs-B (based on clinical observations and blood results), was as accurate as expert multidisciplinary clinician judgement5. However, existing models only consider a binary outcome of death from a particular time point e.g. 30 days. An objective model estimating risk of death over a range of time periods including the last days of life is needed.
We do not know how people die from cancer. In the last 2 weeks of life, there is evidence for deranged respiratory and renal function variables6, although few patients have evidence of organ failure. Pulmonary embolus and infection are thought to be the major causes of death based on post-mortem studies7,8. However, it is unusual for people with cancer to die suddenly as anticipated from a pulmonary embolus. About a third of patients with advanced cancer admitted to specialist palliative care units have a femoral deep vein thrombosis. Therefore, thromboembolism is considered a manifestation of advanced disease, rather than a cause of premature death9