OVERVIEW OF STROKE DATA-BANKS

Stroke data banks (SDBs) used for clinical research are between traditional case series and population-based studies. They can serve a multitude of purposes: estimating the burden of disease treated in one or several centers, calculating the recruitment rates for clinical trials, or for the detection of stroke incidence or time trends by means of a long-time observation in a defined geographical area. Some SDBs restrict themselves to the systematic collection of rare cases of stroke or to recurring strokes. The strength of an SDB lies in its ability to accumulate a large set of data within specified time frames using a defined set of diagnostic procedures. One of the most attractive features of an SDB is its generation of research hypotheses that allow one to look for similar patterns in related and in unrelated disorders. The overview in this volume is based on a literature search of SDBs published since 1980 and a questionnaire sent to the principal investigators applying ten criteria considered essential for clinical research: diagnostic criteria, separate analysis of first-ever strokes, CT investigation rate > 70%, integration of autopsy data, collection of prospective data wherever possible, planning of a pilot phase and interrater studies, constant screening procedures to identify patients enrolled and constant time frame of examinations, a large spectrum of clinical and investigative data, 12 months' follow-up examinations, and a baseline paper describing the procedures used. When applying these criteria to the literature only 8 SDBs fulfilled them. The main reasons for exclusion were restriction to either one stroke subtype, one sex or a limited age group, no specification of stroke subtypes other than ischemic or hemorrhagic, follow-up examinations restricted to 1 month case fatalities.