Medical records from 1,117 cases of cerebrovascular disease from a university and four community hospitals contained data adequate to define the care process, case severity and rates of death and disability (outcome) for patients with certain stroke syndromes. Outcome for patients with "acute stroke" and "vague syndrome" attributed to stroke was different according to the patient's hospital and attending physician. The more frequent case fatality among generalists' patients with vague syndrome was related to case severity; generalists' patients were older and more severely obtunded. Different fatality rates for acute stroke according to the patient's attending physician and hospital, and for vague syndrome according to hospital, were not explained by differences in case severity. Instead, these rates correlated with the care used to prevent, diagnose and treat pneumonia. Examination of the care process for these stroke patients also suggested: (a) that the brain scan and electroencephalogram can be used less frequently in some hospitals in cases of acute stroke, (b) that a more careful neurological examination and other diagnostic work is needed for young or comatose patients with vague syndrome, (c) that oral papaverine is used with unnecessary frequency, and (d) that the indications for anticoagulation of stroke patients should be reviewed.