Machine learning for risk assessment and diagnosis of heart failure

Abstract

Congestive heart failure has become a growing public health problem. Reducing the high cost of congestive heart failure is challenging, as it progresses silently for years before diagnosis, especially in people with high cardiovascular risk and who do not control predisposing factors. New technological advances such as artificial intelligence offer solutions to these problems. Therefore, in this narrative review we determine the application of machine learning for risk identification and diagnosis of heart failure. The search was carried out in English and Spanish in the databases PubMed, HINARI, Google Scholar and Elsevier with the following MeSH terms: «Artificial intelligence», «Machine Learning», «Algorithm», «Cardiology», «Heart Failure», «Heart Failure/diagnosis», and «Heart Failure/prevention and control». We considered original articles, metaanalyses, literature reviews, and systematic reviews, including both cases and controls, published within the last seven years. No artificial intelligence was used in the preparation of this document. Artificial intelligence allows for risk assessment of heart failure and facilitates its timely diagnosis through the analysis of cardiac imaging techniques.