To identify all literature regarding depression in patients with advanced cancer and among mixed hospice populations, and to summarise the prevalence of depression according to different definitions. A systematic review was performed using extensive electronic and hand searches. All studies with quantitative data on prevalence of depression were included and categorised according to their definition of depression. We identified 46 eligible studies giving information on the prevalence of depression, and a further four which gave information on case finding. The most widely used assessment of depression was the Hospital Anxiety and Depression Scale (HADS), which gave a median prevalence of 'definite depression' (i.e., a score on the depression subscale of > 10) of 29%, (interquartile range, IQR, 19.50-34.25%). Studies that used psychiatric interviews indicated a prevalence of major depressive disorder ranging from 5% to 26%, with a median of 15%. Studies were generally small (median sample size 88.5, IQR 50-108), had high numbers of nonresponders, and rarely gave confidence intervals for estimates of prevalence. Depression is a common problem in palliative care settings. The quality of much of the available research is poor, based on small samples of patients with very high nonparticipation rates. The clinical importance of depression is described in subsequent papers.