This study compared the views of palliative care reported by patients, informal carers and the Community Care Team (CCT), a multidisciplinary team caring for people with late stage HIV/AIDS illness. Patients and their carers were interviewed at home, 3-4 weeks after referral to CCT. They rated nine items of the Support Team Assessment Schedule (STAS), a standardized measure of palliative care. Items included current problems such as pain and symptom control, anxiety and service needs. Satisfaction with health services was also recorded. CCT separately recorded the severity of 17 STAS items as part of a continuing audit of care. Relatively few patients (19) and carers (8) were interviewed. Main reasons for non-interview of (105) patients were: 57 too ill and 30 less than 4 weeks in care. CCT's audit showed that non-interviewed patients had significantly more severe problems for five out of 17 STAS items. Patients and CCT identified continuing problems with symptom control, pain control, patient and family anxiety, and communication from professionals. Agreement between patient, carer and CCT ratings was reasonable. Patients and CCT ratings were significantly correlated (Spearman rho = 0.66, p < 0.005). However, patients rated pain as significantly more severe than did CCT (p < 0.05, Wilcoxon Z = -2.45). All patients and seven carers rated the care given by CCT as good or excellent. There were negative comments about communication with other professionals. Studies of palliative care which rely on data gained by patient interview may be biased to include patients with fewer problems. To overcome this providers may wish to audit their care. This study indicates that the views of palliative teams are a reasonable reflection of patients' and carers' experiences, and that the STAS is a valid tool, which we hope will be useful for those wishing to audit their work.