Development & Deployment of a Real-time Healthcare Predictive Analytics Platform

Area of Science:

• Health Informatics
• Clinical Decision Support Systems
• Machine Learning in Healthcare

Background:

• Integrating predictive analytics into healthcare workflows is challenging.
• Electronic Health Record (EHR) data holds significant potential for clinical insights.
• Existing systems often lack scalability and seamless integration capabilities.

Purpose of the Study:

• To present a scalable, cloud-based, fault-tolerant platform for processing EHR data.
• To enable real-time data extraction and result transfer within healthcare systems.
• To facilitate the deployment of predictive analytic algorithms.

Main Methods:

• Development of a cloud-based platform for EHR data processing.
• Utilization of interoperable data standards for portability.
• Integration with UC San Diego Health system for deployment.

Main Results:

• The platform successfully extracts and processes EHR data for patients post-admission.
• Results are transferred back into the EHR system.
• The platform is operational and supports a deep learning model for sepsis prediction.

Conclusions:

• The developed platform offers a scalable and fault-tolerant solution for EHR data integration.
• It enables seamless deployment of predictive analytics in clinical settings.
• The system is clinically relevant, currently used for early sepsis prediction in emergency departments.