SIVA: a hybrid knowledge-and-model-based advisory system for intensive care ventilators

Area of Science:

• Biomedical Engineering
• Artificial Intelligence in Medicine
• Critical Care Medicine

Background:

• Mechanical ventilation is crucial in intensive care but requires complex management.
• Intensive care ventilator management demands adaptive, patient-specific decision support.
• Existing systems may lack the adaptability and robustness needed for dynamic critical care scenarios.

Purpose of the Study:

• To develop and evaluate the Sheffield Intelligent Ventilator Advisor (SIVA) for intensive care ventilator management.
• To create a hybrid system combining fuzzy rule-based and model-based approaches for adaptive decision support.
• To assess the system's performance, appropriateness of advice, and robustness under simulated clinical conditions.

Main Methods:

• Developed a hybrid advisory system with a fuzzy rule-based top-level module and a model-based lower-level module.
• Structured the system for adaptive, patient-specific decision support in invasive and noninvasive modes.
• Validated the top-level module advice against retrospective data and simulated various clinical scenarios with noise and disturbances using closed-loop simulations.

Main Results:

• The Sheffield Intelligent Ventilator Advisor demonstrated appropriate advice in closed-loop simulations.
• Blood gases resulting from the system's decision support were found to be acceptable.
• The system exhibited tolerance to noise and disturbances, indicating robustness in dynamic conditions.

Conclusions:

• The hybrid Sheffield Intelligent Ventilator Advisor offers effective adaptive decision support for intensive care ventilator management.
• The system's performance in simulations suggests its potential utility in real-world clinical settings.
• The validated approach provides a robust and reliable tool for critical care clinicians.