Real-time peri-operative data integration and clinical decision support: an open interoperable platform for patient...

Area of Science:

• Biomedical Informatics
• Health Data Science
• Clinical Engineering

Background:

• Modern ICUs and ORs generate vast physiological data, often underutilized due to interoperability issues and fragmented data.
• Lack of standardized continuous physiological time-series data hinders multicentric collaboration and AI model development.

Purpose of the Study:

• To develop an open, interoperable peri-operative data platform for integrating and utilizing fragmented physiological data.
• To create a scalable, vendor-neutral infrastructure for real-time clinical applications and long-term research.

Main Methods:

• Developed a GDPR-compliant platform using Apache Kafka for real-time data streaming and hybrid long-term storage.
• Integrated data from diverse sources including patient monitors, ventilators, EEG, and lab values.
• Ensured data governance through on-premise storage, anonymization, and ethical approvals.

Main Results:

• Recorded data from 118,915 patients across 77 OR/ICU beds since March 2019.
• Database includes diverse clinical specialties and comprehensive data types (waveforms, numerics, ECG, EEG, ventilation, etc.).
• Collected extensive data on cardiac output, ventilation, ECG analysis, EEG, and syringe pump usage.

Conclusions:

• Demonstrated feasibility of a scalable, vendor-neutral peri-operative data infrastructure.
• The platform supports secure, cross-institutional collaboration and preserves data integrity.
• Provides a reproducible blueprint for interoperable data platforms, advancing precision medicine and clinical equity.