The Use of Artificial Intelligence (AI) to Predict Heart Failure in Type II Diabetes Mellitus Patients: A Systematic Review

Abstract

Heart failure (HF) is a critical concern for individuals with Type II Diabetes Mellitus (T2DM), significantly increasing morbidity and mortality rates. Artificial Intelligence (AI) and machine learning hold promise in enhancing predictive capabilities and guiding personalised interventions. This systematic review evaluates existing AI models' effectiveness in predicting HF complications in T2DM patients. A comprehensive literature search identified 8 relevant studies, predominantly from European, North American, and Southeastern populations. These studies utilised multi-centered registries and electronic medical records to develop AI models predominantly focused on supervised learning algorithms. While the AI models had promising performance, these models lack external validation with diverse populations and reproducibility, hindering their clinical applicability. Moreover, variations in outcome definitions and input features underscore the need for standardised approaches. Despite these limitations, AI models offer valuable insights into HF risk assessment in T2DM, highlighting the importance of further validation and reproducibility for clinical integration.

Keywords

• Artificial Intelligence, Heart Failure, Type 2 Diabetes Mellitus, Machine Learning