Development of Cardiovascular Disease Prediction and Patient Management System Using AI in The Gambia. A Comparative Study of Machine Learning Models

Abstract

Cardiovascular diseases (CVDs) pose a global health challenge, particularly in low-income countries like The Gambia. This study investigates the efficacy of machine learning algorithms in predicting CVDs and explores IoT technology and an AI web application to enhance disease prediction and patient management. Leveraging historical patient data from the Edward Francis Small Teaching Hospital in Banjul (EFSTH), The Gambia, the dataset comprises 915 rows and 9 columns, encompassing diverse demographics and medical profiles. The combined RF, LR, and SVM algorithms achieve 94.5% and 95% accuracy during validation and testing, respectively. Results include the secure SONART AI Cardiovascular Prediction System, offering precise predictions, patient monitoring, and secure data storage. The system upholds confidentiality through encryption and access controls, serving as a cutting-edge healthcare solution in resource-constrained settings. This research contributes valuable insights into predicting CVDs in The Gambia, enhancing patient outcomes and healthcare technologies. IoT integration, utilizing Arduino medical sensors, further empowers disease prediction and patient care. In conclusion, this thesis addresses accurate CVD prediction and early detection needs in The Gambia. By leveraging machine learning and IoT, it advances health informatics, fostering innovative approaches in healthcare delivery. The findings exemplify a commitment to technology and data security, elevating healthcare standards in resource-limited regions.