The Cardiovascular Care 4.0: Predictive Prevention in Primary Health Care through Machine Learning Models

Abstract

Introduction: Cardiovascular risk assessment (CVA) in primary health care (PHC) faces limitations when based solely on traditional scores. Artificial Intelligence (AI) emerges as a promising alternative by increasing predictive accuracy, personalizing care, and potentially reducing health inequalities, especially in vulnerable populations. However, its adoption is still hampered by the lack of transparency of models and the scarcity of solid evidence on performance, safety, and clinical impact. Objective: To evaluate the effectiveness, limitations, and practical applications of AI in predicting and managing CVR in PHC. Methods: An Integrative Literature Review (ILR) was conducted, guided by the question: “How can artificial intelligence be used in primary care for the management of patients with cardiovascular diseases?”. The search was conducted in the Virtual Health Library, Cochrane Library, Directory of Open Access Journals, Elton B. Stephens Company, Publication Medical, Network of Scientific Journals from Latin America and the Caribbean, Spain, and Portugal, and Scientific Electronic Library Online, using the Health Sciences Descriptors: “Artificial Intelligence,” “Primary Health Care,” and “Cardiovascular Diseases” in Portuguese, English, and Spanish. A total of 53 articles were identified; after applying inclusion and exclusion criteria and removing duplicates, 11 articles comprised the final sample. Results: Studies show that AI has great potential in PHC, especially in CVR stratification. Predictive models have shown high accuracy in the early identification of higher-risk individuals and can assist in clinical referral and therapeutic planning. Despite this, important limitations remain, such as a lack of robust evidence on safety, algorithmic biases, model transparency, and real impact on relevant clinical outcomes, such as mortality and reduction of cardiovascular events. Conclusion: AI has advanced significantly and can improve diagnostic accuracy, support clinical decisions, and identify vulnerable individuals in PHC. However, its effective incorporation depends on greater methodological rigor, validation in real-world settings, and ensuring equity in the use of technologies.