Background Research suggests that most people would prefer to die at home, yet patients with haematological malignancies (leukaemia, lymphoma and myeloma), frequently die in hospital; more so than people with other conditions. Although anecdotal accounts exist, little evidence is available to explain this phenomenon. The aim of this study was to explore beliefs about why hospital deaths predominate.

Methods We conducted a qualitative interview study within the Haematological Malignancy Research Network (www.hmrn.org). Participants included 46 clinicians who have contact with blood cancer patients and 10 relatives of deceased patients. Interviews were audio-taped, transcribed and analysed for thematic content. Data coding and classification were carried out using an inductive approach which included: familiarisation with the data; attribution of data to individual codes; collating codes into themes; and interpretation by seeking meaning, salience and connexions.

Results A rich dataset was generated with overlapping themes. Key findings include differing perspectives about hospital deaths, and a ‘mismatch’ between the beliefs of haematology clinicians and those of other interviewees. Haematologists questioned “the rhetoric that the only good death is a home death’, whilst others considered home death preferable. There was recognition that preferences often change and should not be regarded as static. Hospital admission and death were often precipitated by sudden deterioration, distressing symptoms, or unrealistic expectations about home death and the support available to facilitate this. In contrast to the steady decline predicted in oncology patients, the blood cancer pathway was considered unpredictable, with prolonged, toxic treatments, and a fluctuating, uncertain trajectory that may “dip up and down”. This blurred decision-making around treatment intent and identification of the optimal time to discuss this.

Conclusion A complex interplay of factors influenced whether patients with haematological malignancies achieve their preferred place death. Expressed preferences should not be regarded as static, but rather shifting over time, and requiring frequent reassessment, particularly in the context of acute deterioration. Community support services should be adequately resourced if patients are to make meaningful choices that can be delivered. Close collaborations across different health settings are considered fundamental to achieving preferences. Death at home may, but does not necessarily, equate to a ‘good’ death; but while some patients choose to die in hospital, many others die there who have different preferences. Provision of high quality, hospital-based end-of-life care is essential if patient/caregiver experiences are to be optimised.