[A method for predicting probability of stroke]. / Metodika prognozirovaniia veroiatnosti vozniknoveniia insul'ta.

Objectives. To develop a mathematical model of stroke probability in the individual patient and social groups. Material and methods. A large amount of laboratory research was performed in a hospital of Karaganda city. Correlations between the event "stroke"and 35 risk factors were determined. A regression model "stroke-risk factors"was developed using logistic regression method. Results and сonclusions. The equations obtained in the study allow to predict the probability of stroke. Since the regression equation is inherently close to the determined one, it can be differentiated. First derivative of the probability of each of the risk factors allowed us to determine the rate of increase of probability. This provision is new.