Artificial intelligence application in the prevention of chronic non-communicable diseases: a systematic review of publications from 2022 to 2025

Chronic non-communicable diseases (CNCDs) account for 74% of all causes of death worldwide. The development of artificial intelligence (AI) and machine learning technologies (ML/DL) opens up opportunities for a shift from the reactive model of healthcare to proactive and personalized medicine. Objective. To analyze application of artificial intelligence in diagnosis, risk assessment and personalization of prevention of major chronic non-communicable diseases (diabetes mellitus, cardiovascular diseases, chronic respiratory pathology, obesity and metabolic syndrome) according to the literature (2022—2025). Materials and methods. The search was done in PubMed, Scopus and Web of Science. The final analysis includes 102 publications that meet the PRISMA quality criteria out of 3442 articles. Results. High efficiency of AI technologies in risk stratification has been proven. The ML/DL models exceed traditional scales in predicting diabetes mellitus (accuracy >91%, AUC 0.93 for gestational diabetes), cardiovascular events (C-statistic 0.773 compared to 0.759, p<0.0001), chronic obstructive pulmonary disease (accuracy >80%) and metabolic syndrome (AUC 0.85—0.95). Digital twins and analysis of wearable devices allow to personalize the interventions on lifestyle modification. Conclusion. Artificial intelligence has a great potential to transform preventive medicine. However, widespread implementation is limited by problems of model validation, algorithmic bias and the need for data protection. The application of reporting standards (TRIPOD-AI) and interdisciplinary cooperation are needed.