Adaptive API for Real-Time Streaming Analytics as a Service.

Area of Science:

• Biomedical Engineering
• Health Informatics
• Data Science

Background:

• Neonatal intensive care units (NICUs) generate vast amounts of real-time physiological data from bedside monitors.
• Integrating this high-velocity data into analytical processes for real-time insights presents a significant scalability challenge.

Purpose of the Study:

• To demonstrate an adaptive streaming application program interface (API) for real-time data ingestion and distribution.
• To enable multiple analytical services to consume and process live data streams from NICU bedside monitors.

Main Methods:

• Designed and developed an adaptive API for multiple data ingestion from bedside monitors.
• Implemented a middleware for data standardization and structuring.
• Distributed standardized data streams as a service for multiple analytical applications.

Main Results:

• The adaptive API facilitates real-time data processing for individual patients and cohort analysis.
• The system demonstrates robust scalability, adapting to changing data volumes and sources.
• The Artemis Platform successfully instantiated the architecture for high-speed physiological data collection.

Conclusions:

• The proposed adaptive streaming API offers a scalable and robust solution for real-time data analysis in NICUs.
• This architecture supports diverse analytical needs, from single-patient monitoring to multi-patient evaluation.
• The automated process minimizes manual intervention and enhances the efficiency of knowledge discovery from live data streams.