Epidemiology of stroke: methods and trends

Abstract

Epidemiology is the study of the natural history of disease or injury. For a given disorder, this includes its definition, frequency, severity, course and risk factors. Only population-based rates provide data that can be compared among studies. These rates are the incidence or attack rate, the mortality or death rate, and the prevalence rate. For stroke even these rates are not comparable unless they are age-adjusted to a common base. Risk factors are attributes associated with the occurrence of the disease. Relative risk calculation requires population-based rates or a population cohort followed prospectively. For retrospective case-control comparisons the odds ratio is used as an approximation of relative risk. Mortality rates for stroke are widely available and can give a first estimate of risk by geographic location, age, sex and race. In death data, only the sum of all stroke deaths is sufficiently valid for use, while the components are more often wrong than right, except for subarachnoid hemorrhage. International death rates show considerable variation, the highest rates occurring in the Orient. Rates in many countries declined between the 1950s and 1970s but in others, like Portugal, they increased. Stroke deaths in the United States hae been declining rather steadily since the 1920s and continue to do so to the present.