Cardiovascular risk prediction in the family health strategy

A review of machine learning algorithms

Abstract

Cardiovascular disease remains the leading cause of global morbidity and mortality, necessitating robust primary prevention strategies adapted to the reality of health systems. This study aims to conduct a systematic literature review on the application of Machine Learning algorithms for cardiovascular risk prediction within the context of Primary Health Care, specifically in the Family Health Strategy. The methodology employed consisted of a narrative and qualitative review conducted in the PubMed database, covering publications between January 2023 and December 2026, focusing on the comparison between artificial intelligence models and traditional risk scores. The results indicate that machine learning algorithms, such as Random Forest, XGBoost, and Deep Learning, consistently outperform conventional statistical methods in accuracy and risk discrimination, in addition to enabling the use of low-cost digital biomarkers and the integration of social determinants and mental health variables into predictive models. It is concluded that the implementation of these technologies, combined with telemedicine, presents a cost-effective and promising strategy for personalized care and resource optimization in the Unified Health System