This article analyzes the possibilities of strengthening preventive measures based on the analysis of large amounts of data in medicine. Based on clinical and statistical data of 75 thousand patients in the healthcare system of Uzbekistan, predictive models were developed for early detection of cardiovascular diseases, diabetes and obesity risks.The possibility of high-accuracy prediction based on the Random Forest algorithm was proven. The results of the study allow for the personalization of medical prevention and optimization of resources.