To obtain useful results from clinical trials requires attention to various aspects of the trial design. Important among these is the selection of appropriate endpoints or outcome events. Currently measured endpoints involve determination of impairment, disability or handicap, but these represent three distinct levels of outcome. The importance of psychosocial factors such as social support is being increasingly appreciated as a factor contributing to stroke recovery. In addition, attempts are currently being made to provide suitable measures of the patient''s quality of life. There is a considerable lack of clarity and consistency about the meaning and measurement of this variable. Details are given of a new attempt to produce a reliable scoring system to measure quality of life using the patient''s own ranking of important items. Intermediate (surrogate) endpoints can be of value but only if they are also clinically relevant. Finally, the question of sample size is important for the design of clinical trials. Many published trials have contained too few patients to exclude the possibility of some undetected therapeutic benefit. Stroke trials of the future would seek to combine a range of endpoints, including biochemistry and clinical observation with assessment of life quality based on the reports of patients themselves.