In a controlled trial of a home-care service available for the first 6 months after acute stroke, 440 patients received the new service and 417 patients were in the control group. The trial group used more hospital bed days, had a slightly higher admission rate, and did not show better emotional adjustment to stroke than the control group. There was no difference between the 2 groups in stress on relatives. Functional recovery was equal in the 2 groups. A quarter of patients managed at home in each group were severely disabled. Providing a new service does not necessarily alter clinical decisions in the short term, and care should be taken before expanding domiciliary services to reduce hospital use.