Objective: The closure of a long-stay psychiatric hospital in Sydney caused the transfer of an initial 40 very long-term patients to four community residences, each with 10 beds, for a continuing process of deinstitutionalisation. Community psychiatric service support and 24-h supervision were provided. This paper describes the residents' clinical progress which was assessed over a 2-year period. Method: This study employed a quasi-experimental longitudinal design. Evaluation commenced prior to discharge and continued for 2 years following community relocation using the Brief Psychiatric Rating Scale, Life Skills Profile, Social Behaviour Scale, Montgomery Asberg Depression Rating Scale and Quality Of Life measures. Readmission, demographic, case history and medication data were also collected. Results: Of the 40 patients initially transferred to the community, seven required long-term readmission to hospital (either prior to or after amalgamation) and one patient died of medical causes. Additional patients transferred from the hospital to the community following the readmissions. Three of these additional patients had achieved a 2-year community tenure during the study period and were included in the clinical evaluation. The 35 residents in total who remained in the community for 2 years, demonstrated a significant improvement in psychotic symptoms, without significant change in the level of neuroleptic medication. Importantly, the 2 years of community living resulted in a significant increase in the residents' life satisfaction. There were no statistically significant changes in residents' living skills, depressive symptoms or social behaviour problems over the 2 years, indicative of the need for supervision and community service support following deinstitutionalisation. Over the 2-year period, some 37% of the residents required temporary readmission. Conclusion: This study demonstrates the clinical effectiveness of deinstitutionalisation, when planned within a mental health system with adequate community resources.
