Design and Implementation of an Intensive Care Unit Command Center for Medical Data Fusion.

Area of Science:

• Healthcare Technology
• Artificial Intelligence
• Internet of Things

Background:

• The global population is aging, with projections indicating an aging society by 2050.
• Intensive Care Units (ICUs) face high medical error rates, a leading cause of mortality.
• ICU patients require complex care, increasing vulnerability to medical errors.

Purpose of the Study:

• To introduce an innovative AIoT-enabled data fusion system for the CMUH Respiratory Intensive Care Unit (RICU).
• To address and mitigate medical errors in ICUs through advanced data management.
• To enhance patient outcomes and operational efficiency in critical care settings.

Main Methods:

• Implementation of a four-layer AIoT architecture for real-time and non-real-time medical data management.
• Utilizing a three-node streaming cluster (Kafka) for uninterrupted RICU operations.
• Applying an AI application for acute respiratory distress syndrome (ARDS) diagnosis.

Main Results:

• The system efficiently handles 22 TB of medical data annually with minimal delay (1.72 ms) and high bandwidth (65.66 Mbps).
• Ensured uninterrupted RICU operation with a robust streaming cluster.
• AIoT data fusion improved ARDS diagnosis from 52.2% to 93.3% and reduced mortality from 56.5% to 39.5%.

Conclusions:

• AIoT technology offers significant potential for improving patient care and operational efficiency in ICUs.
• The developed AIoT data fusion system demonstrates a viable solution for reducing medical errors and mortality.
• This approach is crucial for managing complex healthcare data and enhancing diagnostic accuracy in critical care.