Services for reducing duration of hospital care for acute stroke patients.

BACKGROUND: Stroke patients conventionally receive a substantial part of their rehabilitation in hospital. Services have now been developed which offer patients in hospital an early discharge with community-based rehabilitation (early supported discharge, ESD). OBJECTIVES: The objective of this review is to establish the costs and effects of ESD services compared with conventional services. SEARCH STRATEGY: The Cochrane Stroke Group Specialised Register was searched and supplemented with information from individual trialists. Searching was completed in March 1999. SELECTION CRITERIA: Randomised controlled trials recruiting stroke patients in hospital to receive either conventional care or any intervention which has provided rehabilitation and support in a community setting with an aim of reducing the duration of hospital care. DATA COLLECTION AND ANALYSIS: Two reviewers scrutinised trials and categorised them on their eligibility. Standardised information was then obtained from the primary trialists. Results were analysed for all trials and for two subgroups depending on whether the intervention was provided by a coordinated multidisciplinary team (coordinated ESD team) or not. MAIN RESULTS: Outcome data are currently available for four trials. Patients tended to be a selected elderly group with disability. Overall, the odds ratio (95% confidence interval) for death, death or institutionalisation, death or dependency at the end of scheduled follow up were 0.87 (0.39-1.93), 0.69 (0.36-1.31) and 0.88 (0.49-1. 57) respectively. Apparent benefits more evident in the three trials evaluating a coordinated ESD team. The ESD group showed significant reductions (P<0.001) in the length of hospital stay equivalent to approximately nine days. REVIEWER'S CONCLUSIONS: ESD services provided for a selected group of stroke patients can reduce the length of hospital stay. However, the relative risks and benefits and overall costs of such services remain unclear.