Background Many trials in the USA, Canada, Europe and Australia have attempted to evaluate the effectiveness of preventive in-home health assessment and home care programs for older people. Trials have differed widely in their processes, including the dependence levels of subjects, assessment components and locations (clinic/home), intensity of case management (frequency of contact, length of follow-up period, scope of interventions) and methods of case management (telephone/visits). Preventive programs use valuable health resources and, although there has been inconclusive evidence of their effectiveness, programs combining preventive in-home health assessment and home care for older people have been introduced into public policy in Australia and internationally. Ongoing research is therefore essential in order to identify the positive benefits for older people, and establish their effectiveness with regard to health resource utilisation. Purpose The purpose of the study was to maintain the health status of older people living alone in the community by implementing a preventive health assessment and follow-up home care program. Research Design An experimental group was compared with a control group using a cluster randomised controlled trial methodology. Health outcomes were measured pre and post intervention, including health perception, functional ability, psychosocial status, client satisfaction, and health resource utilisation. Population and Sample The population for this study consisted of people aged 75 years and over who lived alone in Independent Living Units within managed retirement facilities, and who were highly independent in their activities of daily living. The final sample totalled 124, comprising of an experimental group (n=61) and a control group (n=63). The sample resided in South East Queensland. Intervention The intervention for the study "A Community Preventive Health Model for over 75s Living Alone" comprised of five major elements: 1) targeting before health and/or social crisis, and while community care needs were low; 2) linking clients with a community nurse; 3) comprehensive health assessments and identification of needs; 4) introduction of basic health care and community services and referrals if required; and 5) case management by three-monthly telephone contact. Assessments and case management were carried out by experienced community care registered nurses, and case management was performed for a one year period. The control group received health assessments and phone calls similar to the experimental group for data collection purposes, and to balance the risk of a Hawthorne effect due to regular contacts with participants. However all aspects of case management were omitted from all episodes of contact with the control group. For ethical reasons control group participants were supplied with a summary of their health assessment results to share with their GP if they wished. Data Collection and Instruments Measures of health perception, functional ability and psychosocial status occurred at two points (baseline and after 12 months). Measures of health resource utilisation, mortality and client satisfaction were measured after twelve months. A combination of several widely-used, valid and reliable instruments, as well as some newly developed data collection tools, were used to measure health outcomes. Data Analysis Independent group t-tests and Chi-square tests were used to examine for baseline differences between the experimental and control groups, and also to analyse health resource utilisation data at Time 2. A series of ANCOVA tests were applied to test the remaining hypotheses, so that the effects of Time 1 scores and potential confounding variables could be incorporated into the analyses. Results The experimental group and control group were homogenous at baseline for all demographic variables and all major outcome variables. The intervention model was applied for one year, with 66% (n = 40) in the experimental group having at least one unmet need identified and appropriate interventions undertaken. Only a small proportion of interventions (16%) were recorded as not being followed through by clients, and the majority (59%) resulted in needs being met or problems resolved. Results showed no benefits were gained from the program after one year for experimental group participants for the outcomes of health perception, functional ability, psychosocial status, health resource utilisation and mortality. However, the experimental group did show a statistically significantly higher level of satisfaction with care. Conclusions Comprehensive assessments performed by Registered Nurses with expertise in gerontology resulted in the identification of previously undetected unmet needs. When comprehensive assessment was combined with low intensity case management for a one year period, higher levels of client satisfaction with care were achieved. Therefore a model involving Registered Nurses with advanced knowledge and experience in aged care, working in collaboration with General Practitioners and community service organisations, could have considerable benefits in identifying unmet needs and improving client satisfaction. However, no client benefit was detected for quality of life outcomes, nor was a reduction in health resource utilisation found. This result from an Australian cohort is consistent with findings from many other international trials (Van Haastregt et al., 2000). It is possible that methodological issues are masking the effect of the intervention. Are we measuring appropriate outcomes? Are we expecting long-term outcomes in short time frames? Are we applying the model appropriately across a diverse older population? Further research to explore these questions is recommended for the future.