Predictive analytics and AI hold great promise for healthcare but require careful integration to bridge the gap between ML engineers and healthcare providers. While Electronic Health Record (EHR) adoption is widespread, achieving interoperability across diverse healthcare settings such as long-term care, behavioral health, and public health remains challenging. Integrating heterogeneous data sources into EHRs provides a complete view of patient health, enhancing healthcare delivery through AI-powered intelligent systems. Timely detection and precise diagnosis of heart failure, leveraging diverse datasets and ML techniques, can improve treatment outcomes, reduce charges, and improve patient care. Health AI systems benefit from the rich information contained in clinical narratives and medical imaging, strengthening diagnostic and prognostic models. Data fusion techniques integrate multimodal data, offering a holistic view of patients’ conditions and enhancing clinical decision-making. Our study employs joint fusion and convolutional neural networks (CNNs) to integrate cardiac imaging features with EHR data. Following preprocessing, we train and validate a Support Vector Machine (SVM) model to accurately predict heart disease risk.