BACKGROUND Despite palliative care being an accepted role of community hospitals, there is little quantitative evidence of the type of care provided. OBJECTIVE To obtain quantitative data comparing palliative cancer care provided in 12 community hospitals in 10 towns (approximately 350 medical beds) and in a consultant-led purpose-built hospice (12 beds). METHODS Retrospective medical and nursing case note analysis over one year of cancer deaths in the former Exeter Health District. RESULTS A total of 171 community hospital and 116 hospice casenotes were analysed. Hospice patients had significantly different reasons for admission compared with community hospital patients (P < 0.001), with pain and symptom control being more frequent and terminal nursing care less frequent reasons for admission to the hospice. Community hospital length of stay was significantly longer than hospice length of stay (P = 0.002; mean community hospital stay 16 days, mean hospice stay eight days). Symptoms on admission differed significantly. Drug prescribing on admission and at death and indications of active treatment of symptoms were broadly similar. Community hospital patients received more investigations than hospice patients, linked to the observation that around one in ten community hospital patients were admitted for investigation and active treatment. Community hospital medical notes were significantly less likely to meet minimum quality standards than were hospice notes (81/171 vs. 18/116; P < 0.001), with major deficiencies in the areas of examination, progress reporting, and absence of confirmation of death. CONCLUSIONS This study confirms the role of community hospitals in palliative terminal cancer care. Differences in care between community hospitals and a hospice have been demonstrated that may reflect either different admission populations to each setting or differences in the way care was delivered.