A programme of regular planned respite admissions to the geriatric unit in Cork was established in 1984, for dependent elderly being cared for at home. Retrospective analysis of the respite programme sought to establish its bed utilisation, type of patients admitted, factors in their selection process, their outcome and any associated hospital morbidity. Twenty-four patients have used the scheme during the four years since its commencement, with an efficiency of bed utilisation in 1987 of 14 patients occupying a mean of 4.33 beds per month. The median number of medical problems per patients was five; their median age 77 years and the percentage over 80 years was 46%. Twenty-two patients were already receiving intensive rehabilitation before being placed on the programme, mobility problems dominated. Nine patients continue on the programme with a mean of 23.9 months of respites to date, eight were discharged and seven died. There was no significant difference between the number of respite stays in hospital and "at home" stays complicated by morbidity. The results suggest that even for a highly selected group of dependent elderly, a planned programme of respite admissions is an effective form of care, uses hospital beds efficiently without associated morbidity and may provide an alternative to long stay institutional care.