The use of modern digital technologies in predictive analysis of risk factors for premature death due to socially significant non-communicable diseases (literature review)

Abstract

The effectiveness of the implementation of the Concept of predictive, preventive and personalized medicine is directly related to the development and scaling of the process of digitalization of healthcare with the leading position occupied by artificial intelligence technologies (AI technologies). This fully applies to the problem of predictive analysis of risk factors for premature death from socially significant non-communicable diseases (NCDs). The purpose of the work was to summarize the current domestic and foreign experience of using AI technologies and machine learning (ML) in predictive analysis of risk factors for premature death from socially significant non-communicable diseases. The search for publications was carried out in the RSCI, CyberLeninka, eLibrary, and PubMed databases containing domestic and foreign sources of scientific information. The search depth covered period from 2011 to 2021. More than 50 sources of scientific information were analyzed. The article briefly reports on the global risk factors (RF) of premature death due to NCDs, the main place among which is occupied by diseases of the circulatory system. The disadvantages of calculators used in mass examinations to determine the total risk of fatal cardiovascular events (CVE) are considered ¾ Framingham scale and SCORE scale. It is shown that the individual predictive efficiency of calculators can be increased due to ML technologies that use big data on the health status of the population in certain regions, digitalization of medical images, and expansion of structured databases of the RF spectrum, which makes it possible to recognize and take into account complex relationships between multiple, correlated, and nonlinear RF and CVE outcomes. Examples of the predictive effectiveness of ML models are given. Special attention is paid to AI technologies and deep ML in the stratification of CVE risk and outcomes based on the analysis of imagesof the fundus the eye. Conclusion. The introduction of AI technologies and ML in clinical practice opens up the prospect of achieving an effective individualized stratification of the risk of premature death due to chronic NCDs and their factor of personalized prevention through timely optimization of socially significant diseases modifiable by the F.