Factors associated with hospital death in patients with haematological malignancies: findings from a UK population-based study

    loading  Checking for direct PDF access through Ovid



Research suggests most people would prefer to die at home, yet patients with haematological malignancies (leukaemia, lymphoma and myeloma) die in hospital more often than people with other conditions, though reasons for this are unclear.2,3


To investigate factors associated with hospital deaths in patients with blood cancers.


This study is set within the Haematological Malignancy Research Network (www.HMRN.org),4 an ongoing population-based cohort that follows patients from diagnosis onwards. All 886 adult patients diagnosed 2004–2012 and dying 2011–2012 were included, with data collected from their medical records.


Over half (56.9%) the 886 patients died in hospital. After adjusting for sex and age, several additional factors remained associated with the likelihood of hospital death (P < 0.05). Patients receiving treatment with life-prolonging/curative intent in the 30 days before death were more likely to have died in hospital compared to those receiving palliative care (OR 3.83, 95% CI: 2.40–6.11). Those with documented discussions about preferred place of care/death were around half as likely to have died in hospital as those without discussions (OR 0.46, 95% CI: 0.35–0.61). Final discussions not involving the patient themselves were also associated with an increased likelihood of hospital death (family only vs patient involvement, OR 2.31 95% CI: 1.50–3.57).


Hospital deaths were most likely to occur in patients receiving life-prolonging/curative treatment, suggesting death was not anticipated; and when the patient themselves was not involved in discussions, implying advanced illness. This indicates that hospital deaths are often unexpected or unavoidable and that maximising the quality of end-of-life care in hospitals should be a priority.

Related Topics

    loading  Loading Related Articles