Usefulness of a Population-Based Stroke Registry

Abstract

Stroke registries can supply eminent informations concerning epidemiological, clinical and socio-economical data providing they are reliable, i.e. compliant with Malmgren''s criteria and the Oxford registration model. The collection must include a well-defined population. It must be exhaustive, specific and prospective in order to prevent omissions and to facilitate the elimination of other diagnoses. Only the first stroke has to be taken into account for each patient and the clinical reports have to be complete (usefulness of a simplified but flexible work file). Finally, the rate of diagnoses either doubtful or based on death certificate has to be low. A CT scan should be performed for each patient and a maximum number of information sources is required. Consequently, a stroke registry (1) provides descriptive epidemiological stroke studies. They are based on the estimation of incidence and mechanisms of stroke. Their main value is to appreciate the extent of the health problem, to forecast the expected number of patients and to assess the quality of data collection by controlling their consistency from one year to the next. These studies can determine the distribution of stroke among the 3 traditional health care procedures (hospitals, private clinics and home care). This information is of major importance when undertaking clinical or prevention trials. Finally, the collection of survival data from the studies is important to state the natural history of stroke and therefore to appreciate their alteration following a special treatment or new preventive procedures. (2) A stroke registry allows some contribution to clinical research by providing the distribution ratio of different stroke subtypes and mechanisms, the clinical data for semiological analyses and the adjusted prognosis criteria according to data from nonhospitalized cases. (3) It also allows the development of analytic epidemiological studies. Such a registry avoids some bias. Analysis of etiological factors is done on case-control and cohort studies. These registries are of great methodological help to the analyses of drug trials. (4) The registry leads to experimental epidemiological studies. It is a good tool for controlling health care quality and confirming the role of a factor underscored by an analytic epidemiological study: the consequences of the risk factor reduction or eviction (e.g. smoking, alcohol, high blood pressure, dyslipidemia) provide additional evidence towards its etiological role. (5) It provides sanitary information. The diffusion of these results to clinicians, searchers and sanitary authorities is of great importance. All these advantages do not abolish the relevance of hospital registries, which despite recruitment bias, provide much more accurate clinical studies, CT scans in almost 100% of the patients, and a better follow-up, such as in Lausanne.