In a study of nursing home applicants, residential care decisions made by a multidisciplinary assessment team were examined. The team agreed that of the 296 applicants assessed, 54 per cent required the high level of physical and supportive care provided in a nursing home. Hostel care was recommended for 17 per cent, continuing care at home for 17 per cent, hospice care for 1 subject, and in 13 per cent of cases the team postponed their decision. A decision for nursing home care was associated with low Barthel Index of Activities of Daily Living scores, dementia, incontinence and the absence of a carer willing to continue care. A decision to delay was associated, in most cases, with a requirement for further in-patient assessment and/or rehabilitation, and therefore with a potential for functional improvement. A decision for hostel care instead of home care was associated with a low level of informal support and the absence of a carer who was a spouse or daughter. The findings suggest that a program of geriatric assessment will accurately identify the differing care needs of nursing home applicants. The initial value of such an assessment program may be to contribute to the planning of residential and other long-term care services rather than to reduce inappropriate nursing home admissions. (Awt J Public Health 1991; 15: 222-7) benefit of geriatric assessment programs is a reduction in the number of nursing home A admissions required by dependent elderly people. The evidence for this has been summarised in a review of available studies and includes a number of randomised controlled trials of in-patient geriatric assessment units.' The issue of whether or not assessment programs should be specifically targeted at nursing home applicants to prevent inappropriate nursing home admissions is less clear. Descriptive studies involving assessment of nursing home applicants do suggest that in the absence of adequate assessment, treatment and/or rehabilitation, elderly people are being admitted to nursing homes when they could be cared for in less supportive settings.2-6 A similar situation has been reported in the United Kingdom for long-term hospital care.' One British study, however, found only a 7 per cent discrepancy between the application for residential care and the assessed need for residential care and suggested that this was the real case as the sample included all applicants in the area.' Also, apart from descriptive