Background and Purpose: Stroke rehabilitation units have been shown to improve mortality and reduce morbidity and hospital length of stay as compared with conventional medical ward care. In contrast, the effectiveness of acute stroke units (ASUs), which only provide early in-patient care, is unknown. We have compared ASU care with that administered on a general medical or geriatric ward. Methods: Historical comparison of 116 consecutive patients admitted to a new ASU during the 6-month period from February to July 1994 with 128 consecutive patients admitted to general medical or geriatric wards during the previous 6-month period (August 1993 to January 1994). Results: Admission baseline characteristics were similar between the two groups. As compared with conventional general ward care, patients admitted to the ASU had a shorter stay in casualty, 6 h (4–6) versus 7 h (4–11; 2p = 0.03); reduced length of stay in hospital, 20 days (10–55) versus 31 days (13–80; 2p = 0.09); increased carotid Doppler investigation, 40/94 versus 25/101 (2p = 0.013); increased prescription of secondary prevention measures, aspirin 49/94 versus 33/101 (2p = 0.009) or treatment of hypertension 19/50 versus 5/47 (2p = 0.004); improved outcome, home: institution: death, 69:11:18 versus 63:13:32 (2p = 0.049). Conclusion: ASUs appear to contribute to improved care (investigations, secondary prevention) and outcome following stroke whilst patients spend less time in h