A large number of patients die from heart failure. While a small proportion of deaths in severe heart failure are sudden, the majority will die from worsening heart failure or a comorbid condition. A retrospective study of patients with heart disease in the UK, and a prospective study of heart failure in the USA, showed that a high proportion of patients had uncontrolled symptoms at the end of life.1 ,2 In addition there was evidence of communication problems between health care professionals and patients and their carers, in particular, open communication about dying and patients' preferences about how and where they should be cared for.3Significant communication problems as well as unmet psychosocial needs have also been identified in a prospective study.4

Specialist palliative care has traditionally offered a multiprofessional approach to manage such problems in patients with cancer and more recently motor neurone disease and AIDS. Its potential role in improving care for patients with heart failure at the end of life has been recognised in strategic plans for health care in the UK.5 Nevertheless there is very limited information about the management of heart failure patients by specialist palliative care. We conducted a retrospective study of the needs and service use of heart failure referrals to specialist palliative care.

St Christopher's Hospice is a specialist palliative care unit in London with 62 inpatient beds and over 450 community patients covering a population of 1.75 million. Since 1994 referrals of patients with heart disease have been accepted. We identified these cases from a comprehensive database of all referrals between 1994 and 1999 and reviewed the case notes.

Twenty seven patients were referred because of heart failure out of a total of 9920. The mean age of heart failure patients was 73 years (range 48–98 years). Fifteen referrals were made from primary care, 5 from a hospital palliative care team, 5 from a hospital doctor, and 2 from a non-St Christopher's community palliative care team. Four patients died before assessment (delay 1–4 days). Four patients were not accepted (2 were inappropriate on assessment and required help from psychiatric and social services, 1 was out of area, and for 1 there were inadequate data).

Thus 19 patients were assessed and managed by the hospice. Eleven patients had significant comorbidity. The most common symptoms reported as severe on initial assessment were weakness (11 patients), breathlessness (6), drowsiness (5), anorexia (5), dry mouth (3) and weight loss (3).

The non-physical needs recognised on initial assessment were adjustment of the patient and family to dying (7 patients), exhaustion of carers (5), psychological distress of the family (5), communication and relationship problems between patient and family or carers (4), adjustment difficulties of the family associated with fear of loss and guilt (4), anxiety and frustration of patient (3), and miscellaneous spiritual, financial, and insight issues.

The specialist palliative care team delivered care both in the community and as hospice day and inpatients. It provided help with recognition and management of psychological distress of family or carers (7 patients), discontinuing unnecessary cardiac and non-cardiac drugs (7), early bereavement follow up of family or carers at risk (7), mobilising increased nursing, social or financial support (4), and appropriate medication required for terminal care (4). Adjustments to cardiac medication were made in consultation with the patients' physicians. Palliative prescribing included commencing 6 patients on morphine and 2 on benzodiazepines for dyspnoea, and 2 patients on diazepam for anxiety. Two patients underwent abdominal paracentesis for relief of uncomfortable ascites.

Seven patients required inpatient care only, 8 home care only, and 4 both. The median time of inpatient care was 15 days (range 1–63 days; all hospice admissions mean 15.6 days), and of home care was 25 days (range 1–867 days; all hospice patients mean 109.8 days).

Three patients were discharged before death. Of the remaining 16 patients, 10 died in the hospice and 6 at home.

This study shows that selected patients with heart failure can be managed by specialist palliative care which can address both inadequate symptom control and psychosocial and communication problems by drawing on its large experience in cancer. Patients with heart disease made similar demands on the specialist palliative care service when compared to other hospice patients. The majority of patients died in the hospice, a change in the place of death from standard care where a similar number died in hospital.3 This study cannot determine whether specialist palliative care improved outcomes or indeed which interventions are appropriate in this group of patients. The role of specialist palliative care in providing services for patients with heart failure is yet to be determined.