Home care of the elderly is of increasing concern not only to purchasers and providers of health care but also to the public and to those responsible for providing social care. As with any service, the aim must be to provide care that is appropriate for each individual. To achieve that, valid and reliable measures of a person's needs are required and resources are to be used as efficiently as possible. A considerable amount of work has been carried out to develop such normative-based measures for assessing the home care needs of the elderly in the form of comprehensive geriatric assessment (CGA). CGA is a commonly used technology which has been shown to be associated with improved health status and lower service use. Despite widespread use, however, the effectiveness of different CGAs has not yet been fully investigated. In the Province of Quebec, Canada, two CGAs which differ in comprehensiveness and resource requirements are being used to assess needs at entry to home care. The aim of this study is to compare the differential impact of these two CGA procedures on patient outcomes: the Systdme de mesure de l'autonomie fonctionnelle, the longer, more comprehensive and resourceintensive CGA, and the Admission au maintien d domicile which is a shorter and less resource-intensive form of CGA. In a prospective cohort study, 158 elderly patients aged 65 years or over were assessed at admission to home care using one or the other CGA and changes in health-related quality of life as well as service use were monitored and compared at the end of a 12-week follow-up. Costs related to the use of a long or a shortform CGA were also explored. These comparisons were made while controlling for patient (age, gender, living alone, quality of life at entry, depression), process (type and intensity of care received) and structural variables (budget and staff mix). Results from comparative and multivariate analyses are in favour of not rejecting the null hypothesis that both forms of CGAs are similarly associated with outcomes. Depression was the strongest predictor of changes in quality of life and high intensity of care and a low proportion of nurses on the home care teams were the strongest predictors of service use outside HC. These results lead us to discuss whether long or shortform CGAs were developed on a comparative rather than a normative definition of needs. The implications of these findings for home care policy and practice are discussed and suggestions for future research are presented.