E-SIMS: A Human-Centered Vision for Personalized Elderly Health Monitoring

Abstract

In 2024, approximately 4.1 million Americans turned 65, joining 18% of the current population that is 65+. Our current healthcare system, especially for the elderly, is largely reactive, with treatment often beginning only after symptoms appear and are reported by the patient. This can lead to delays in treatment and poorer outcomes, as patients are often unreliable narrators. While technologies like smartwatches and medical sensors are widely used, they typically function in isolation and provide fragmented information. Furthermore, most AI-based health tools tend to focus on one-size-fits-all models that fail to reflect individual patient differences, treating patients with complex health conditions the same as healthy individuals. This lack of personalization, combined with limited healthcare resources, makes early detection and timely intervention difficult. To address these limitations, this paper presents a vision for Elder Symptoms Individualized Monitoring Systems (E-SIMS), a human-centered reference architecture for proactive elderly health monitoring. The proposed framework integrates multimodal health data, edge intelligence, federated learning, and explainable artificial intelligence to support personalized risk assessment and early intervention workflows. By combining edge-cloud collaboration with clinician-in-the-loop validation and privacy-aware communication mechanisms, E-SIMS aims to enable scalable, trustworthy, and individualized monitoring for aging populations. We outline the conceptual architecture of the system and identify key research challenges that must be addressed to realize this vision in real-world healthcare environments.