ABSTRACT
On the basis of the data on 2287 patients with cerebral strokes (CS) and 800 individuals without this disease, and using discriminant and probability methods, the authors have identified 158 supposed risk factors (RF) of CS development. They have also established the specificity of RFs which depends on the nature of CS, as well as ascertained the dependence of informative value of these factors on the size of samples and the degree of percentage reflection by the main demographic indicators of the studied population. RFs are classified as controllable and uncontrollable. The authors have selected 50 risk factors with 2-5 gradations ranging from the normal to the maximal pathology and reflecting various aspects of CS etiopathogenesis. This work has resulted in the creating of a mathematical model of SC development. On the basis of this model and using the Bayes' theorem the authors have developed a system intended for predicting CS and its nature by 50, 25, and 15 RFs which is implemented by means of universal, micro- or minicomputers and tables which make it possible to predict with a 88.8% accuracy the development of CS as well as to determine the most rational individual prophylaxis of threatening stroke which is achieved through an accurate balancing in the system of controllable and uncontrollable RFs.
