Background and Purpose-We have previously shown that treatment of acute stroke patients in our stroke unit (SU) compared with treatment in general ward (GWs) improves short-and long-term survival and functional outcome and increases the possibility of earlier discharge to home. The aim of the present study was to identify the differences in treatment between the SU and the GW and to assess which aspects of the SU care which were most responsible for the better outcome. Methods-Of the 220 patients included in our trial, only 206 were actually treated (SU, 102 patients; GW, 104 patients).For these patients, we identified the differences in the treatment and the consequences of the treatment. We analyzed the factors that we were able to measure and their association with the outcome, discharge to home within 6 weeks. Results-Characteristic features in our SU were teamwork, staff education, functional training, and integrated physiotherapy and nursing. Other treatment factors significantly different in the SU from the GW were shorter time to start of the systematic mobilization/training and increased use of oxygen, heparin, intravenous saline solutions, and antipyretics. Consequences of the treatment seem to be less variation in diastolic and systolic blood pressure (BP), avoiding the lowest diastolic BP, and lowering the levels of glucose and temperature in the SU group compared with the GW group. Univariate analyses showed that all these factors except the level of glucose were significantly associated with discharge to home within 6 weeks. In the final multivariate Cox regression model, shorter time to start of the mobilization/training and stabilized diastolic BP were independent factors significantly associated with discharge to home within 6 weeks. Conclusions-Shorter time to start of mobilization/training was the most important factor associated with discharge to home, followed by stabilized diastolic BP, indicating that these factors probably were important in the SU treatment. The effects of characteristic features of an SU, such as a specially trained staff, teamwork, and involvement of relatives, were not possible to measure. Such factors might be more important than those actually measured. (Stroke. 1999;30:917-923.)