A Fog-Based AI Doctor Copilot for Real-Time Clinical Risk Triage in Telehealth IoT Systems

Abstract

The increasing adoption of telehealth and remote patient monitoring has created new challenges for clinical decision support, particularly when timely responses are required for large numbers of geographically distributed patients. Rather than replacing clinical judgment, current artificial intelligence (AI) applications are increasingly designed to support healthcare professionals by identifying patients at elevated risk, prioritizing alerts, and assisting clinical workflows. Building on this direction, this study presents a fog-based AI Doctor Copilot for real-time clinical risk triage in Internet of Things-enabled telehealth environments. The proposed framework integrates continuous physiological monitoring with adaptive machine learning models deployed on fog computing nodes, enabling patient risk to be dynamically classified as low, medium, or high while preserving clinician oversight throughout the decision-making process. System performance was assessed through simulation using synthetic physiological data, including heart rate, blood pressure, and blood glucose measurements, and benchmarked against a conventional cloud-based clinical decision support system. Compared with cloud-only architecture, the proposed approach reduced alert latency by as much as 52%, decreased false alerts by approximately 31%, and lowered overall energy consumption by about 22%, while maintaining comparable risk stratification accuracy. These findings suggest that distributing AI inference to the fog layer can improve the responsiveness and efficiency of telehealth decision support without compromising clinical oversight, making the approach well suited to time-sensitive remote care environments. Because the evaluation was conducted using synthetic patient data in a simulation environment, the results should be interpreted as evidence of architectural feasibility rather than clinical effectiveness.

Received: 8 May 2026 | Revised: 12 July 2026 | Accepted: 24 July 2026

Conflicts of Interest

The authors declare that they have no conflicts of interest to this work.

Data Availability Statement

The data that support this work are available upon reasonable request to the corresponding author.

Author Contribution Statement

Nathan Guo: Methodology, Data curation, Writing – review & editing, Visualization. Bryan Guo: Software, Writing – review & editing, Visualization. Yunyong Guo: Conceptualization, Methodology, Validation, Formal analysis, Investigation, Writing – original draft, Writing – review & editing, Project administration. Alex Kuo: Conceptualization, Validation, Resources, Data curation, Writing – review & editing, Supervision.