The distinction between factors that influence survival after stroke and those that increase the likelihood of institutionalization is an important health issue. Estimates of survival and frequency of institutionalization after stroke vary widely, depending on the patient population. A precise picture of variability of outcome from stroke may be obtained from a prospective epidemiologic study. This report uses the Framingham Study population sample of 5,184 men and women, aged 30 to 62 at entry in 1948, who were free of cardiovascular disease. All completed strokes that occurred between 1971 and 1981 were evaluated. Of the 213 patients with completed strokes, 154 survived more than 30 days. Multivariate logistic regression analysis indicated that acute survival was negatively influenced by stroke type, severity of neurologic impairment, and age. For those who survived at least 30 days, independent living was determined by social factors as much as by severity of disability. Being married protected men but not women from institutionalization. Older women, married or not, with moderate to severe residual impairment and minimal education, were at highest risk of institutionalization. In acute stroke, medical factors dominated rates of survival. However, in those who survived, family and social factors had an equal impact in determining final outcome from stroke.