This chapter explores the evolving landscape of healthcare informatics, focusing on the transition from data collection and organization to the conversion of multimodal information into actionable knowledge. Our vision is a healthcare informatics system in which validated models within EHRs offer real-time risk assessments and personalized patient care suggestions. Such a system will allow clinicians to combine advanced data-driven insights with conventional clinical judgment, leading to personalized and more effective patient management. However, in order to accomplish this objective, it is essential to employ sophisticated multimodal machine learning techniques, meticulously collect and annotate patient data, strengthen data engineering infrastructures, establish strategic implementation strategies, and conduct real-world trials across several centers. This chapter will provide a brief overview of the current state of contemporary EHRs for healthcare informatics, followed by concise evaluations of current health informatics research focusing on common chronic diseases like diabetes, heart disease, cancer, neurological disorders, and mental health, with a specific emphasis on a spectrum of approaches (unimodal to multimodal) to extract actionable knowledge from healthcare data.