To assess the magnitude of the problem of stroke, data were summarised from several national surveys conducted by the National Center for Health Statistics. This report describes the occurrence of stroke and the utilisation of healthcare services for stroke in the US, with special attention to the population 65 years of age and over. In 1993, 150 108 deaths were attributed to stroke, the third leading cause of death. Death rates rose steeply with age and men had higher death rates than women. Death rates were lower in Whites than in Blacks with the exception of ages 85 and over. Considerable geographic variation in stroke mortality was also noted.Since 1979, stroke death rates have continued the long term decline, albeit slowed, in each age, sex and race group. There were probably nearly 3 million persons with diagnosed stroke in the civilian noninstitutionalised population in the US in 1994. In 1993, over 841 000 persons were discharged from short-stay hospitals with a principal diagnosis of stroke. The discharge rate was 328 per 100 000. The hospital fatality rate was 7.8% for all cerebrovascular disease.Since 1988, hospitalisation rates for stroke have changed little while hospital fatality rates have decreased. In 1994, there were 2 million visits to physicians' offices for stroke, 1.6 million in persons aged 65 and over. Among nursing home residents in the 1980s, the prevalence of diagnosed stroke was 193 per 1000. In 1995, expenditures for stroke totalled $US18 000 million in the US. Although mortality rates from stroke among persons 65 years of age and over have continued to decrease since 1979, disability and hospitalisation rates and utilisation of other services stress the need for even more vigorous efforts for primary prevention of stroke to reduce the burden of illness, and not just the mortality rate from stroke.