Factors affecting the prediction of health care costs in cardiovascular patients using artificial intelligence tools: a systematic review of regression and machine learning approaches

Abstract Introduction Cardiovascular diseases (CVDs) are a public health issue in many countries and the primary cause of disease burden worldwide, leading to the increasing cost of health care. Artificial intelligence (AI) and machine learning (ML) tools offer promising solutions for predicting health care costs for cardiovascular patients, enhancing resource allocation, and improving financial planning in health care systems. It is very important to identify factors that can help predict the treatment costs of cardiovascular patients. Therefore, the purpose of our systematic review was to identify the factors affecting the prediction of health care costs in cardiovascular patients using artificial intelligence tools. Methods PubMed, Scopus, Web of Science, IEEE Xplore, and Google Scholar databases were searched from inception until 21 March 2025. Eligibility criteria included original studies without restriction in study design, studies published in the English language, and studies that have reported effective factors in predicting the health costs of cardiovascular patients. The quality of the included studies was assessed by using the Effective Public Health Practice Project quality assessment tool (EPHPP). Results A total of 2,189 records were identified, of which 1,772 unique records remained after removing 417 duplicates. After screening, 36 studies met the inclusion criteria. The most significant factors influencing health care costs were age, gender, type of cardiovascular procedure (e.g., Coronary Artery Bypass Grafting (CABG), Percutaneous Transluminal Coronary Angioplasty (PTCA)), comorbidities, length of hospitalization, and complications (e.g., respiratory failure, infection, kidney failure). AI models, such as logistic regression, decision trees, random forests, and neural networks, were commonly used to predict costs. However, only a small number of studies applied machine learning-based approaches, indicating that the use of ML techniques in this field is still limited. Conclusion Machine learning models showed superior accuracy in some studies compared with traditional regression methods for predicting healthcare costs in cardiovascular patients; however, this evidence was limited to a small number of studies and was not consistent across all datasets. Recent advancements in technologies such as AI and ML models can significantly improve the prediction of healthcare costs. By identifying the key factors that influence treatment costs and using advanced predictive models, the accuracy of cost estimations can be enhanced. This will enable health managers and policymakers to make more informed decisions regarding resource allocation and financial planning.